conference
stringclasses 6
values | title
stringlengths 2
176
| abstract
stringlengths 2
5k
| decision
stringclasses 11
values |
|---|---|---|---|
ICLR.cc/2021/Conference | Nearest Neighbor Machine Translation | We introduce $k$-nearest-neighbor machine translation ($k$NN-MT), which predicts tokens with a nearest-neighbor classifier over a large datastore of cached examples, using representations from a neural translation model for similarity search. This approach requires no additional training and scales to give the decoder direct access to billions of examples at test time, resulting in a highly expressive model that consistently improves performance across many settings. Simply adding nearest-neighbor search improves a state-of-the-art German-English translation model by 1.5 BLEU. $k$NN-MT allows a single model to be adapted to diverse domains by using a domain-specific datastore, improving results by an average of 9.2 BLEU over zero-shot transfer, and achieving new state-of-the-art results---without training on these domains. A massively multilingual model can also be specialized for particular language pairs, with improvements of 3 BLEU for translating from English into German and Chinese. Qualitatively, $k$NN-MT is easily interpretable; it combines source and target context to retrieve highly relevant examples. | Accept (Poster) |
ICLR.cc/2022/Conference | Understanding ResNet from a Discrete Dynamical System Perspective | Residual network (ResNet) is one of popular networks proposed in recently years. Discussion about its theoretical properties is helpful for the understanding of networks with convolution modules. In this paper, we formulate the learning process of ResNet as a iterative system, then we may apply tools in discrete dynamical systems to explain its stability and accuracy. Due to the backward propagation of learning process, the module operations vary with the change of different layers. So we introduce the condition number of modules to describe the perturbation of output data, which can demonstrate the robustness of ResNet. In addition, the inter-class and intra-class median principal angle is defined to analyze the classification efficiency of ResNet. Mathematical description of the learning process of ResNet is given in a modular manner so that our research framework can be applied to other networks. In order to verify the feasibility of our idea, several experiments are carried out on the Dogs vs. Cats dataset, Kaggle: Animals 10 dataset, and ImageNet 2012 dataset. Simulation results are accordance with the theoretical analysis and prove the validity of our theory. | Withdrawn |
ICLR.cc/2021/Conference | HiFiSinger: Towards High-Fidelity Neural Singing Voice Synthesis | High-fidelity singing voices usually require higher sampling rate (e.g., 48kHz, compared with 16kHz or 24kHz in speaking voices) with large range of frequency to convey rich expression and emotion. However, higher sampling rate results in wider frequency band and longer waveform sequence with more fine-grained details and presents challenges for singing modeling in both frequency and time domains in singing voice synthesis (SVS). In this paper, we develop HiFiSinger, an SVS system towards high-fidelity singing voice using 48kHz sampling rate. HiFiSinger consists of a FastSpeech based neural acoustic model and a Parallel WaveGAN based neural vocoder to ensure fast training and inference and also high voice quality. To tackle the difficulty of singing modeling caused by high sampling rate (wider frequency band and longer waveform), we introduce multi-scale adversarial training in both the acoustic model and vocoder to improve singing modeling. Specifically, 1) To handle the larger range of frequencies caused by higher sampling rate (e.g., 48kHz vs. 24kHz), we introduce a novel sub-frequency GAN (SF-GAN) on mel-spectrogram generation, which splits the full 80-dimensional mel-frequency into multiple sub-bands (e.g. low, middle and high frequency bands) and models each sub-band with a separate discriminator. 2) To model longer waveform sequences caused by higher sampling rate, we introduce a multi-length GAN (ML-GAN) for waveform generation to model different lengths of waveform sequences with separate discriminators. 3) We also introduce several additional designs in HiFiSinger that are crucial for high-fidelity voices, such as adding F0 (pitch) and V/UV (voiced/unvoiced flag) as acoustic features, choosing an appropriate window and hop size for mel-spectrogram, and increasing the receptive field in vocoder for long vowel modeling in singing voices. Experiment results show that HiFiSinger synthesizes high-fidelity singing voices with much higher quality: 0.32/0.44 MOS gain over 48kHz/24kHz baseline and 0.83 MOS gain over previous SVS systems. Audio samples are available at https://hifisinger.github.io. | Withdrawn |
ICLR.cc/2019/Conference | NECST: Neural Joint Source-Channel Coding | For reliable transmission across a noisy communication channel, classical results from information theory show that it is asymptotically optimal to separate out the source and channel coding processes. However, this decomposition can fall short in the finite bit-length regime, as it requires non-trivial tuning of hand-crafted codes and assumes infinite computational power for decoding. In this work, we propose Neural Error Correcting and Source Trimming (NECST) codes to jointly learn the encoding and decoding processes in an end-to-end fashion. By adding noise into the latent codes to simulate the channel during training, we learn to both compress and error-correct given a fixed bit-length and computational budget. We obtain codes that are not only competitive against several capacity-approaching channel codes, but also learn useful robust representations of the data for downstream tasks such as classification. Finally, we learn an extremely fast neural decoder, yielding almost an order of magnitude in speedup compared to standard decoding methods based on iterative belief propagation. | Reject |
ICLR.cc/2021/Conference | Information-Theoretic Odometry Learning | In this paper, we propose a unified information-theoretic framework for odometry learning, a crucial component of many robotics and vision tasks such as navigation and virtual reality where 6-DOF poses are required in real time. We formulate this problem as optimizing a variational information bottleneck objective function, which eliminates pose-irrelevant information from the latent representation. The proposed framework provides an elegant tool for performance evaluation and understanding in information-theoretic language. Specifically, we bound the generalization errors of the deep information bottleneck framework and the predictability of the latent representation. These provide not only a performance guarantee but also practical guidance for model design, sample collection, and sensor selection. Furthermore, the stochastic latent representation provides a natural uncertainty measure without the needs for extra structures or computations. Experiments on two well-known odometry datasets demonstrate the effectiveness of our method. | Withdrawn |
ICLR.cc/2021/Conference | Leveraged Weighted Loss For Partial Label Learning | As an important branch of weakly supervised learning, partial label learning deals with data where each instance is assigned with a set of candidate labels, whereas only one of them is true. In this paper, we propose a family of loss functions named Leveraged Weighted (LW) loss function, which for the first time introduces the leverage parameter $\beta$ to partial loss functions to leverage between losses on partial labels and residual labels (non-partial labels). Under mild assumptions, we achieve the relationship between the partial loss function and its corresponding ordinary loss that leads to the consistency in risk. Compared to the existing literatures, our result applies to both deterministic and stochastic scenarios, considers the loss functions of a more general form, and takes milder assumptions on the distribution of the partial label set. As special cases, with $\beta = 1$ and $\beta = 2$, the corresponding ordinary losses of our LW loss respectively match the binary classification loss and the \textit{one-versus-all} (OVA) loss function. In this way, our theorems successfully explain the experimental results on parameter analysis, where $\beta = 1$ and especially $\beta = 2$ are considered as preferred choices for the leverage parameter $\beta$. Last but not least, real data comparisons show the high effectiveness of our LW loss over other state-of-the-art partial label learning algorithms. | Withdrawn |
ICLR.cc/2021/Conference | Learning Robust Models using the Principle of Independent Causal Mechanisms | Standard supervised learning breaks down under data distribution shift. However, the principle of independent causal mechanisms (ICM, Peters et al. (2017)) can turn this weakness into an opportunity: one can take advantage of distribution shift between different environments during training in order to obtain more robust models. We propose a new gradient-based learning framework whose objective function is derived from the ICM principle. We show theoretically and experimentally that neural networks trained in this framework focus on relations remaining invariant across environments and ignore unstable ones. Moreover, we prove that the recovered stable relations correspond to the true causal mechanisms under certain conditions. In both regression and classification, the resulting models generalize well to unseen scenarios where traditionally trained models fail. | Reject |
ICLR.cc/2023/Conference | Graph Contrastive Learning with Reinforced Augmentation | Graph contrastive learning (GCL), designing contrastive objectives to learn embeddings from augmented graphs, has become a prevailing method for learning embeddings from graphs in an unsupervised manner. As an important procedure in GCL, graph data augmentation (GDA) directly affects the model performance on the downstream task. Currently, there are three types of GDA strategies: trial-and-error, precomputed method, and adversarial method. However, these strategies ignore the connection between the two consecutive augmentation results because GDA is regarded as an independent process. In this paper, we regard the GDA in GCL as a Markov decision process. Based on this point, we propose a reinforced method, i.e., the fourth type of GDA strategy, using a novel Graph Advantage Actor-Critic (GA2C) model for GCL. On 23 graph datasets, the experimental results verify that GA2C outperforms the SOTA GCL models on a series of downstream tasks such as graph classification, node classification, and link prediction. | Reject |
ICLR.cc/2023/Conference | Towards Better Selective Classification | We tackle the problem of Selective Classification where the objective is to achieve the best performance on a predetermined ratio (coverage) of the dataset. Recent state-of-the-art selective methods come with architectural changes either via introducing a separate selection head or an extra abstention logit. In this paper, we challenge the aforementioned methods. The results suggest that the superior performance of state-of-the-art methods is owed to training a more generalizable classifier rather than their proposed selection mechanisms. We argue that the best performing selection mechanism should instead be rooted in the classifier itself. Our proposed selection strategy uses the classification scores and achieves better results by a significant margin, consistently, across all coverages and all datasets, without any added compute cost. Furthermore, inspired by semi-supervised learning, we propose an entropy-based regularizer that improves the performance of selective classification methods. Our proposed selection mechanism with the proposed entropy-based regularizer achieves new state-of-the-art results. | Accept: poster |
ICLR.cc/2021/Conference | Client Selection in Federated Learning: Convergence Analysis and Power-of-Choice Selection Strategies | Federated learning is a distributed optimization paradigm that enables a large number of resource-limited client nodes to cooperatively train a model without data sharing. Several works have analyzed the convergence of federated learning by accounting of data heterogeneity, communication and computation limitations, and partial client participation. However, they assume unbiased client participation, where clients are selected at random or in proportion of their data sizes. In this paper, we present the first convergence analysis of federated optimization for biased client selection strategies, and quantify how the selection bias affects convergence speed. We reveal that biasing client selection towards clients with higher local loss achieves faster error convergence. Using this insight, we propose Power-of-Choice, a communication- and computation-efficient client selection framework that can flexibly span the trade-off between convergence speed and solution bias. We also propose an extension of Power-of-Choice that is able to maintain convergence speed improvement while diminishing the selection skew. Our experiments demonstrate that Power-of-Choice strategies converge up to 3 $\times$ faster and give $10$% higher test accuracy than the baseline random selection. | Reject |
ICLR.cc/2021/Conference | Learned Threshold Pruning | This paper presents a novel differentiable method for unstructured weight pruning of deep neural networks. Our learned-threshold pruning (LTP) method learns per-layer thresholds via gradient descent, unlike conventional methods where they are set as input. Making thresholds trainable also makes LTP computationally efficient, hence scalable to deeper networks. For example, it takes $30$ epochs for LTP to prune ResNet50 on ImageNet by a factor of $9.1$. This is in contrast to other methods that search for per-layer thresholds via a computationally intensive iterative pruning and fine-tuning process. Additionally, with a novel differentiable $L_0$ regularization, LTP is able to operate effectively on architectures with batch-normalization. This is important since $L_1$ and $L_2$ penalties lose their regularizing effect in networks with batch-normalization. Finally, LTP generates a trail of progressively sparser networks from which the desired pruned network can be picked based on sparsity and performance requirements. These features allow LTP to achieve competitive compression rates on ImageNet networks such as AlexNet ($26.4\times$ compression with $79.1\%$ Top-5 accuracy) and ResNet50 ($9.1\times$ compression with $92.0\%$ Top-5 accuracy). We also show that LTP effectively prunes modern \textit{compact} architectures, such as EfficientNet, MobileNetV2 and MixNet. | Reject |
ICLR.cc/2019/Conference | LEARNING ADVERSARIAL EXAMPLES WITH RIEMANNIAN GEOMETRY | Adversarial examples, referred to as augmented data points generated by imperceptible perturbation of input samples, have recently drawn much attention. Well-crafted adversarial examples may even mislead state-of-the-art deep models to make wrong predictions easily. To alleviate this problem, many studies focus on investigating how adversarial examples can be generated and/or resisted. All the existing work handles this problem in the Euclidean space, which may however be unable to describe data geometry. In this paper, we propose a generalized framework that addresses the learning problem of adversarial examples with Riemannian geometry. Specifically, we define the local coordinate systems on Riemannian manifold, develop a novel model called Adversarial Training with Riemannian Manifold, and design a series of theory that manages to learn the adversarial examples in the Riemannian space feasibly and efficiently. The proposed work is important in that (1) it is a generalized learning methodology since Riemmanian manifold space would be degraded to the Euclidean space in a special case; (2) it is the first work to tackle the adversarial example problem tractably through the perspective of geometry; (3) from the perspective of geometry, our method leads to the steepest direction of the loss function. We also provide a series of theory showing that our proposed method can truly find the decent direction for the loss function with a comparable computational time against traditional adversarial methods. Finally, the proposed framework demonstrates superior performance to the traditional counterpart methods on benchmark data including MNIST, CIFAR-10 and SVHN. | Reject |
ICLR.cc/2019/Conference | A PRIVACY-PRESERVING IMAGE CLASSIFICATION FRAMEWORK WITH A LEARNABLE OBFUSCATOR | Real world images often contain large amounts of private / sensitive information that should be carefully protected without reducing their utilities. In this paper, we propose a privacy-preserving deep learning framework with a learnable ob- fuscator for the image classification task. Our framework consists of three mod- els: learnable obfuscator, classifier and reconstructor. The learnable obfuscator is used to remove the sensitive information in the images and extract the feature maps from them. The reconstructor plays the role as an attacker, which tries to recover the image from the feature maps extracted by the obfuscator. In order to best protect users’ privacy in images, we design an adversarial training methodol- ogy for our framework to optimize the obfuscator. Through extensive evaluations on real world datasets, both the numerical metrics and the visualization results demonstrate that our framework is qualified to protect users’ privacy and achieve a relatively high accuracy on the image classification task. | Withdrawn |
ICLR.cc/2023/Conference | Selective Frequency Network for Image Restoration | Image restoration aims to reconstruct the latent sharp image from its corrupted counterpart. Besides dealing with this long-standing task in the spatial domain, a few approaches seek solutions in the frequency domain in consideration of the large discrepancy between spectra of sharp/degraded image pairs. However, these works commonly utilize transformation tools, e.g., wavelet transform, to split features into several frequency parts, which is not flexible enough to select the most informative frequency component to recover. In this paper, we exploit a multi-branch and content-aware module to decompose features into separate frequency subbands dynamically and locally, and then accentuate the useful ones via channel-wise attention weights. In addition, to handle large-scale degradation blurs, we propose an extremely simple decoupling and modulation module to enlarge the receptive field via global and window-based average pooling. Integrating two developed modules into a U-Net backbone, the proposed Selective Frequency Network (SFNet) performs favorably against state-of-the-art algorithms on five image restoration tasks, including single-image defocus deblurring, image dehazing, image motion deblurring, image desnowing, and image deraining. | Accept: poster |
ICLR.cc/2019/Conference | FFJORD: Free-Form Continuous Dynamics for Scalable Reversible Generative Models | A promising class of generative models maps points from a simple distribution to a complex distribution through an invertible neural network. Likelihood-based training of these models requires restricting their architectures to allow cheap computation of Jacobian determinants. Alternatively, the Jacobian trace can be used if the transformation is specified by an ordinary differential equation. In this paper, we use Hutchinson’s trace estimator to give a scalable unbiased estimate of the log-density. The result is a continuous-time invertible generative model with unbiased density estimation and one-pass sampling, while allowing unrestricted neural network architectures. We demonstrate our approach on high-dimensional density estimation, image generation, and variational inference, achieving the state-of-the-art among exact likelihood methods with efficient sampling. | Accept (Oral) |
ICLR.cc/2022/Conference | Inductive Bias of Multi-Channel Linear Convolutional Networks with Bounded Weight Norm | We provide a function space characterization of the inductive bias resulting from minimizing the $\ell_2$ norm of the weights in multi-channel linear convolutional networks. We define an \textit{induced regularizer} in the function space as the minimum $\ell_2$ norm of weights of a network required to realize a function. For two layer linear convolutional networks with $C$ output channels and kernel size $K$, we show the following: (a) If the inputs to the network have a single channel, the induced regularizer for any $K$ is \textit{independent} of the number of output channels $C$. Furthermore, we derive the regularizer is a norm given by a semidefinite program (SDP). (b) In contrast, for networks with multi-channel inputs, multiple output channels can be necessary to merely realize all matrix-valued linear functions and thus the inductive bias \emph{does} depend on $C$. However, for sufficiently large $C$, the induced regularizer is again given by an SDP that is independent of $C$. In particular, the induced regularizer for $K=1$ and $K=D$ are given in closed form as the nuclear norm and the $\ell_{2,1}$ group-sparse norm, respectively, of the Fourier coefficients.
We investigate the applicability of our theoretical results to a broader scope of ReLU convolutional networks through experiments on MNIST and CIFAR-10 datasets. | Desk_Rejected |
ICLR.cc/2022/Conference | Language Model Pre-training on True Negatives | Discriminative pre-trained language models (PrLMs) learn to predict original texts from intentionally corrupted ones. Taking the former text as positive and the latter as negative samples, the discriminative PrLM can be trained effectively for contextualized representation. However, though the training of such a type of PrLMs highly relies on the quality of the automatically constructed samples, existing PrLMs simply treat all corrupted texts as equal negative without any examination, which actually lets the resulting model inevitably suffer from the false negative issue where training is carried out on wrong data and leads to less efficiency and less robustness in the resulting PrLMs.
Thus in this work, on the basis of defining the false negative issue in discriminative PrLMs that has been ignored for a long time, we design enhanced pre-training methods to counteract false negative predictions and encourage pre-training language models on true negatives, by correcting the harmful gradient updates subject to false negative predictions. Experimental results on GLUE and SQuAD benchmarks show that our counter-false-negative pre-training methods indeed bring about better performance together with stronger robustness. | Reject |
ICLR.cc/2018/Conference | Confidence Scoring Using Whitebox Meta-models with Linear Classifier Probes | We propose a confidence scoring mechanism for multi-layer neural networks based on a paradigm of a base model and a meta-model. The confidence score is learned by the meta-model using features derived from the base model – a deep neural network considered a whitebox. As features, we investigate linear classifier probes inserted between the various layers of the base model and trained using each layer’s intermediate activations. Experiments show that this approach outperforms various baselines in a filtering task, i.e., task of rejecting samples with low confidence. Experimental results are presented using CIFAR-10 and CIFAR-100 dataset with and without added noise exploring various aspects of the method. | Withdrawn |
ICLR.cc/2020/Conference | Semi-supervised Semantic Segmentation using Auxiliary Network | Recently, the convolutional neural networks (CNNs) have shown great success on semantic segmentation task. However, for practical applications such as autonomous driving, the popular supervised learning method faces two challenges: the demand of low computational complexity and the need of huge training dataset accompanied by ground truth. Our focus in this paper is semi-supervised learning. We wish to use both labeled and unlabeled data in the training process. A highly efficient semantic segmentation network is our platform, which achieves high segmentation accuracy at low model size and high inference speed. We propose a semi-supervised learning approach to improve segmentation accuracy by including extra images without labels. While most existing semi-supervised learning methods are designed based on the adversarial learning techniques, we present a new and different approach, which trains an auxiliary CNN network that validates labels (ground-truth) on the unlabeled images. Therefore, in the supervised training phase, both the segmentation network and the auxiliary network are trained using labeled images. Then, in the unsupervised training phase, the unlabeled images are segmented and a subset of image pixels are picked up by the auxiliary network; and then they are used as ground truth to train the segmentation network. Thus, at the end, all dataset images can be used for retraining the segmentation network to improve the segmentation results. We use Cityscapes and CamVid datasets to verify the effectiveness of our semi-supervised scheme, and our experimental results show that it can improve the mean IoU for about 1.2% to 2.9% on the challenging Cityscapes dataset. | Reject |
ICLR.cc/2023/Conference | Push and Pull: Competing Feature-Prototype Interactions Improve Semi-supervised Semantic Segmentation | This paper challenges semi-supervised segmentation with a rethink on the feature-prototype interaction in the classification head. Specifically, we view each weight vector in the classification head as the prototype of a semantic category. The basic practice in the softmax classifier is to pull a feature towards its positive prototype (i.e., the prototype of its class), as well as to push it away from its negative prototypes. In this paper, we focus on the interaction between the feature and its negative prototypes, which is always “pushing” to make them dissimilar. While the pushing-away interaction is necessary, this paper reveals a new mechanism that the contrary interaction of pulling close negative prototypes is also beneficial. We have two insights for this counter-intuitive interaction: 1) some pseudo negative prototypes might actually be positive so that the pulling interaction can help resisting the pseudo-label noises, and 2) some true negative prototypes might contain contextual information that is beneficial. Therefore, we integrate these two competing interactions into a Push-and-Pull Learning (PPL) method. On the one hand, PPL introduces the novel pulling-close interaction between features and negative prototypes with a feature-to-prototype attention. On the other hand, PPL reinforces the original pushing-away interaction with a multi-prototype contrastive learning. While PPL is very simple, experiments show that it substantially improves semi-supervised segmentation and sets a new state of the art. | Reject |
ICLR.cc/2018/Conference | On the Discrimination-Generalization Tradeoff in GANs | Generative adversarial training can be generally understood as minimizing certain moment matching loss defined by a set of discriminator functions, typically neural networks. The discriminator set should be large enough to be able to uniquely identify the true distribution (discriminative), and also be small enough to go beyond memorizing samples (generalizable). In this paper, we show that a discriminator set is guaranteed to be discriminative whenever its linear span is dense in the set of bounded continuous functions. This is a very mild condition satisfied even by neural networks with a single neuron. Further, we develop generalization bounds between the learned distribution and true distribution under different evaluation metrics. When evaluated with neural distance, our bounds show that generalization is guaranteed as long as the discriminator set is small enough, regardless of the size of the generator or hypothesis set. When evaluated with KL divergence, our bound provides an explanation on the counter-intuitive behaviors of testing likelihood in GAN training. Our analysis sheds lights on understanding the practical performance of GANs. | Accept (Poster) |
ICLR.cc/2022/Conference | Explaining Point Processes by Learning Interpretable Temporal Logic Rules | We propose a principled method to learn a set of human-readable logic rules to explain temporal point processes.
We assume that the generative mechanisms underlying the temporal point processes are governed by a set of first-order temporal logic rules, as a compact representation of domain knowledge. Our method formulates the rule discovery process from noisy event data as a maximum likelihood problem, and designs an efficient and tractable branch-and-price algorithm to progressively search for new rules and expand existing rules. The proposed algorithm alternates between the rule generation stage and the rule evaluation stage, and uncovers the most important collection of logic rules within a fixed time limit for both synthetic and real event data. In a real healthcare application, we also had human experts (i.e., doctors) verify the learned temporal logic rules and provide further improvements. These expert-revised interpretable rules lead to a point process model which outperforms previous state-of-the-arts for symptom prediction, both in their occurrence times and types. | Accept (Poster) |
ICLR.cc/2018/Conference | Do Convolutional Neural Networks act as Compositional Nearest Neighbors? | We present a simple approach based on pixel-wise nearest neighbors to understand and interpret the functioning of state-of-the-art neural networks for pixel-level tasks. We aim to understand and uncover the synthesis/prediction mechanisms of state-of-the-art convolutional neural networks. To this end, we primarily analyze the synthesis process of generative models and the prediction mechanism of discriminative models. The main hypothesis of this work is that convolutional neural networks for pixel-level tasks learn a fast compositional nearest neighbor synthesis/prediction function. Our experiments on semantic segmentation and image-to-image translation show qualitative and quantitative evidence supporting this hypothesis. | Withdrawn |
ICLR.cc/2022/Conference | Robust Losses for Learning Value Functions | Most value function learning algorithms in reinforcement learning are based on the mean squared (projected) Bellman error. However, squared errors are known to be sensitive to outliers, both skewing the solution of the objective and resulting in high-magnitude and high-variance gradients. Typical strategies to control these high-magnitude updates in RL involve clipping gradients, clipping rewards, rescaling rewards, and clipping errors. Clipping errors is related to using robust losses, like the Huber loss, but as yet no work explicitly formalizes and derives value learning algorithms with robust losses. In this work, we build on recent insights reformulating squared Bellman errors as a saddlepoint optimization problem, and propose a saddlepoint reformulation for a Huber Bellman error and Absolute Bellman error. We show that the resulting solutions have significantly lower error for certain problems and are otherwise comparable, in terms of both absolute and squared value error. We show that the resulting gradient-based algorithms are more robust, for both prediction and control, with less stepsize sensitivity. | Reject |
ICLR.cc/2022/Conference | Neuron-Enhanced Autoencoder based Collaborative filtering: Theory and Practice | This paper presents a novel recommendation method called neuron-enhanced autoencoder based collaborative filtering (NE-AECF). The method uses an additional neural network to enhance the reconstruction capability of autoencoder. Different from the main neural network implemented in a layer-wise manner, the additional neural network is implemented in an element-wise manner. They are trained simultaneously to construct an enhanced autoencoder of which the activation function in the output layer is learned adaptively to approximate possibly complicated response functions in real data. We provide theoretical analysis for NE-AECF to investigate the generalization ability of autoencoder and deep learning in collaborative filtering. We prove that the element-wise neural network is able to reduce the upper bound of the prediction error for the unknown ratings, the data sparsity is not problematic but useful, and the prediction performance is closely related to the difference between the number of users and the number of items.
Numerical results show that our NE-AECF has promising performance on a few benchmark datasets. | Reject |
ICLR.cc/2018/Conference | Neural Map: Structured Memory for Deep Reinforcement Learning | A critical component to enabling intelligent reasoning in partially observable environments is memory. Despite this importance, Deep Reinforcement Learning (DRL) agents have so far used relatively simple memory architectures, with the main methods to overcome partial observability being either a temporal convolution over the past k frames or an LSTM layer. More recent work (Oh et al., 2016) has went beyond these architectures by using memory networks which can allow more sophisticated addressing schemes over the past k frames. But even these architectures are unsatisfactory due to the reason that they are limited to only remembering information from the last k frames. In this paper, we develop a memory system with an adaptable write operator that is customized to the sorts of 3D environments that DRL agents typically interact with. This architecture, called the Neural Map, uses a spatially structured 2D memory image to learn to store arbitrary information about the environment over long time lags. We demonstrate empirically that the Neural Map surpasses previous DRL memories on a set of challenging 2D and 3D maze environments and show that it is capable of generalizing to environments that were not seen during training. | Accept (Poster) |
ICLR.cc/2023/Conference | On Representing Mixed-Integer Linear Programs by Graph Neural Networks | While Mixed-integer linear programming (MILP) is NP-hard in general, practical MILP has received roughly 100--fold speedup in the past twenty years. Still, many classes of MILPs quickly become unsolvable as their sizes increase, motivating researchers to seek new acceleration techniques for MILPs. With deep learning, they have obtained strong empirical results, and many results were obtained by applying graph neural networks (GNNs) to making decisions in various stages of MILP solution processes. This work discovers a fundamental limitation: there exist feasible and infeasible MILPs that all GNNs will, however, treat equally, indicating GNN's lacking power to express general MILPs. Then, we show that, by restricting the MILPs to unfoldable ones or by adding random features, there exist GNNs that can reliably predict MILP feasibility, optimal objective values, and optimal solutions up to prescribed precision. We conducted small-scale numerical experiments to validate our theoretical findings. | Accept: poster |
ICLR.cc/2018/Conference | Weighted Transformer Network for Machine Translation | State-of-the-art results on neural machine translation often use attentional sequence-to-sequence models with some form of convolution or recursion. Vaswani et. al. (2017) propose a new architecture that avoids recurrence and convolution completely. Instead, it uses only self-attention and feed-forward layers. While the proposed architecture achieves state-of-the-art results on several machine translation tasks, it requires a large number of parameters and training iterations to converge. We propose Weighted Transformer, a Transformer with modified attention layers, that not only outperforms the baseline network in BLEU score but also converges 15-40% faster. Specifically, we replace the multi-head attention by multiple self-attention branches that the model learns to combine during the training process. Our model improves the state-of-the-art performance by 0.5 BLEU points on the WMT 2014 English-to-German translation task and by 0.4 on the English-to-French translation task. | Reject |
ICLR.cc/2022/Conference | Self-Supervised Learning for Binary Networks by Joint Classifier Training | Despite the great success of self-supervised learning with large floating point networks, such networks are not readily deployable to edge devices.
To accelerate deployment of models to edge devices for various downstream tasks by unsupervised representation learning, we propose a self-supervised learning method for binary networks.
In particular, we propose to use a randomly initialized classifier attached to a pretrained floating point feature extractor as targets and jointly train it with a binary network.
For better training of the binary network, we propose a feature similarity loss, a dynamic balancing scheme of loss terms, and modified multi-stage training.
We call our method as BSSL.
Our empirical validations show that BSSL outperforms self-supervised learning baselines for binary networks in various downstream tasks and outperforms supervised pretraining in certain tasks. | Withdrawn |
ICLR.cc/2022/Conference | Learning to Solve Combinatorial Problems via Efficient Exploration | From logistics to the natural sciences, combinatorial optimisation on graphs underpins numerous real-world applications. Reinforcement learning (RL) has shown particular promise in this setting as it can adapt to specific problem structures and does not require pre-solved instances for these, often NP-hard, problems. However, state-of-the-art (SOTA) approaches typically suffer from severe scalability issues, primarily due to their reliance on expensive graph neural networks (GNNs) at each decision step. We introduce ECORD; a novel RL algorithm that alleviates this expense by restricting the GNN to a single pre-processing step, before entering a fast-acting exploratory phase directed by a recurrent unit. Experimentally, we demonstrate that ECORD achieves a new SOTA for RL algorithms on the Maximum Cut problem, whilst also providing orders of magnitude improvement in speed and scalability. Compared to the nearest competitor, ECORD reduces the optimality gap by up to 73% on 500 vertex graphs with a decreased wall-clock time. Moreover, ECORD retains strong performance when generalising to larger graphs with up to 10000 vertices. | Reject |
ICLR.cc/2022/Conference | When high-performing models behave poorly in practice: periodic sampling can help | Training a deep neural network (DNN) for breast cancer detection from medical images suffers from the (hopefully) low prevalence of the pathology.
For a sensible amount of positive cases, images must be collected from numerous places resulting in large heterogeneous datasets with different acquisition devices, populations, cancer incidences.
Without precaution, this heterogeneity may result in a DNN biased by latent variables a priori independent of the pathology.
This may be dramatic if this DNN is used inside a software to help radiologists to detect cancers.
This work mitigates this issue by acting on how mini-batches for Stochastic Gradient Descent (SGD) algorithms are constructed.
The dataset is divided into homogeneous subsets sharing some attributes (\textit{e.g.} acquisition device, source) called Data Segments (DSs).
Batches are built by sampling each DS periodically with a frequency proportional to the rarest label in the DS and by simultaneously preserving an overall balance between positive and negative labels within the batch.
Periodic sampling is compared to balanced sampling (equal amount of labels within a batch, independently of DS) and to balanced sampling within DS (equal amount of labels within a batch and each DS).
We show, on breast cancer prediction from mammography images of various devices and origins, that periodic sampling leads to better generalization than other sampling strategies. | Reject |
ICLR.cc/2022/Conference | Coordinated Attacks Against Federated Learning: A Multi-Agent Reinforcement Learning Approach | We propose a model-based multi-agent reinforcement learning attack framework against federated learning systems. Our framework first approximates the distribution of the clients' aggregated data through cooperative multi-agent coordination. It then learns an attack policy through multi-agent reinforcement learning. Depending on the availability of the server's federated learning configurations, we introduce algorithms for both white-box attacks and black-box attacks. Our attack methods are capable of handling scenarios when the clients' data is independent and identically distributed and when the data is independent but not necessarily identically distributed. We further derive an upper bound on the attacker's performance loss due to inaccurate distribution estimation. Experimental results on real-world datasets demonstrate that the proposed attack framework achieves strong performance even if the server deploys advanced defense mechanisms. Our work sheds light on how to attack federated learning systems through multi-agent coordination. | Withdrawn |
ICLR.cc/2023/Conference | Parameter Averaging for Feature Ranking | Neural Networks are known to be sensitive to initialisation. The methods that rely on neural networks for feature ranking are not robust since they can have variations in their ranking when the model is initialized and trained with different random seeds. In this work, we introduce a novel method based on parameter averaging to estimate accurate and robust feature importance in tabular data setting, referred as XTab. We first initialize and train multiple instances of a shallow network (referred as local masks) with "different random seeds" for a downstream task. We then obtain a global mask model by "averaging the parameters" of local masks. We show that although the parameter averaging might result in a global model with higher loss, it still leads to the discovery of the ground-truth feature importance more consistently than an individual model does. We conduct extensive experiments on a variety of synthetic and real-world data, demonstrating that the XTab can be used to obtain the global feature importance that is not sensitive to sub-optimal model initialisation. | Reject |
ICLR.cc/2023/Conference | Discrete Contrastive Diffusion for Cross-Modal Music and Image Generation | Diffusion probabilistic models (DPMs) have become a popular approach to conditional generation, due to their promising results and support for cross-modal synthesis. A key desideratum in conditional synthesis is to achieve high correspondence between the conditioning input and generated output. Most existing methods learn such relationships implicitly, by incorporating the prior into the variational lower bound. In this work, we take a different route---we explicitly enhance input-output connections by maximizing their mutual information. To this end, we introduce a Conditional Discrete Contrastive Diffusion (CDCD) loss and design two contrastive diffusion mechanisms to effectively incorporate it into the denoising process, combining the diffusion training and contrastive learning for the first time by connecting it with the conventional variational objectives. We demonstrate the efficacy of our approach in evaluations with diverse multimodal conditional synthesis tasks: dance-to-music generation, text-to-image synthesis, as well as class-conditioned image synthesis. On each, we enhance the input-output correspondence and achieve higher or competitive general synthesis quality. Furthermore, the proposed approach improves the convergence of diffusion models, reducing the number of required diffusion steps by more than 35% on two benchmarks, significantly increasing the inference speed. | Accept: poster |
ICLR.cc/2022/Conference | Optimal Representations for Covariate Shift | Machine learning systems often experience a distribution shift between training and testing. In this paper, we introduce a simple variational objective whose optima are exactly the set of all representations on which risk minimizers are guaranteed to be robust to any distribution shift that preserves the Bayes predictor, e.g., covariate shifts. Our objective has two components. First, a representation must remain discriminative for the task, i.e., some predictor must be able to simultaneously minimize the source and target risk. Second, the representation's marginal support needs to be the same across source and target. We make this practical by designing self-supervised objectives that only use unlabelled data and augmentations to train robust representations.
Our objectives give insights into the robustness of CLIP, and further improve CLIP's representations to achieve SOTA results on DomainBed. | Accept (Poster) |
ICLR.cc/2021/Conference | Gradient-based tuning of Hamiltonian Monte Carlo hyperparameters | Hamiltonian Monte Carlo (HMC) is one of the most successful sampling methods in machine learning. However, its performance is significantly affected by the choice of hyperparameter values, which require careful tuning. Existing approaches for automating this task either optimise a proxy for mixing speed or consider the HMC chain as an implicit variational distribution and optimize a tractable lower bound that is too loose to be useful in practice. Instead, we propose to optimize an objective that quantifies directly the speed of convergence to the target distribution. Our objective can be easily optimized using stochastic gradient descent. We evaluate our proposed method and compare to baselines on a variety of problems including synthetic 2D distributions, the posteriors of variational autoencoders and the Boltzmann distribution for molecular configurations of a 22 atom molecule. We find our method is competitive with or improves upon alternative baselines on all problems we consider. | Reject |
ICLR.cc/2018/Conference | Countering Adversarial Images using Input Transformations | This paper investigates strategies that defend against adversarial-example attacks on image-classification systems by transforming the inputs before feeding them to the system. Specifically, we study applying image transformations such as bit-depth reduction, JPEG compression, total variance minimization, and image quilting before feeding the image to a convolutional network classifier. Our experiments on ImageNet show that total variance minimization and image quilting are very effective defenses in practice, in particular, when the network is trained on transformed images. The strength of those defenses lies in their non-differentiable nature and their inherent randomness, which makes it difficult for an adversary to circumvent the defenses. Our best defense eliminates 60% of strong gray-box and 90% of strong black-box attacks by a variety of major attack methods. | Accept (Poster) |
ICLR.cc/2019/Conference | The Cakewalk Method | Combinatorial optimization is a common theme in computer science. While in general such problems are NP-Hard, from a practical point of view, locally optimal solutions can be useful. In some combinatorial problems however, it can be hard to define meaningful solution neighborhoods that connect large portions of the search space, thus hindering methods that search this space directly. We suggest to circumvent such cases by utilizing a policy gradient algorithm that transforms the problem to the continuous domain, and to optimize a new surrogate objective that renders the former as generic stochastic optimizer. This is achieved by producing a surrogate objective whose distribution is fixed and predetermined, thus removing the need to fine-tune various hyper-parameters in a case by case manner. Since we are interested in methods which can successfully recover locally optimal solutions, we use the problem of finding locally maximal cliques as a challenging experimental benchmark, and we report results on a large dataset of graphs that is designed to test clique finding algorithms. Notably, we show in this benchmark that fixing the distribution of the surrogate is key to consistently recovering locally optimal solutions, and that our surrogate objective leads to an algorithm that outperforms other methods we have tested in a number of measures. | Reject |
ICLR.cc/2021/Conference | Generative Time-series Modeling with Fourier Flows | Generating synthetic time-series data is crucial in various application domains, such as medical prognosis, wherein research is hamstrung by the lack of access to data due to concerns over privacy. Most of the recently proposed methods for generating synthetic time-series rely on implicit likelihood modeling using generative adversarial networks (GANs)—but such models can be difficult to train, and may jeopardize privacy by “memorizing” temporal patterns in training data. In this paper, we propose an explicit likelihood model based on a novel class of normalizing flows that view time-series data in the frequency-domain rather than the time-domain. The proposed flow, dubbed a Fourier flow, uses a discrete Fourier transform (DFT) to convert variable-length time-series with arbitrary sampling periods into fixed-length spectral representations, then applies a (data-dependent) spectral filter to the frequency-transformed time-series. We show that, by virtue of the DFT analytic properties, the Jacobian determinants and inverse mapping for the Fourier flow can be computed efficiently in linearithmic time, without imposing explicit structural constraints as in existing flows such as NICE (Dinh et al. (2014)), RealNVP (Dinh et al. (2016)) and GLOW (Kingma & Dhariwal (2018)). Experiments show that Fourier flows perform competitively compared to state-of-the-art baselines. | Accept (Poster) |
ICLR.cc/2023/Conference | Multi-Task Structural Learning using Local Task Similarity induced Neuron Creation and Removal | Multi-task learning has the potential to improve generalization by maximizing positive transfer between tasks while reducing task interference. Fully achieving this potential is hindered by manually designed architectures that remain static throughout training. In contrast, learning in the brain occurs through structural changes that are in tandem with changes in synaptic strength. Therefore, we propose Multi-Task Structural Learning (MTSL) which simultaneously learns the multi-task architecture and its parameters. MTSL begins with an identical single task network for each task and alternates between a task learning phase and a structural learning phase. In the task learning phase, each network specializes in the corresponding task. In each of the structural learning phases, starting from the earliest layer, locally similar task layers first transfer their knowledge to a newly created group layer after which they become redundant and are removed. Our experimental results show that MTSL achieves competitive generalization with various baselines and improves robustness to out-of-distribution data. | Reject |
ICLR.cc/2021/Conference | Mem2Mem: Learning to Summarize Long Texts with Memory Compression and Transfer | We introduce Mem2Mem, a memory-to-memory mechanism for hierarchical recurrent neural network based encoder decoder architectures and we explore its use for abstractive document summarization. Mem2Mem transfers memories via readable/writable external memory modules that augment both the encoder and decoder. Our memory regularization compresses an encoded input article into a more compact set of sentence representations. Most importantly, the memory compression step performs implicit extraction without labels, sidestepping issues with suboptimal ground-truth data and exposure bias of hybrid extractive-abstractive summarization techniques. By allowing the decoder to read/write over the encoded input memory, the model learns to read salient information about the input article while keeping track of what has been generated. Our Mem2Mem approach yields results that are competitive with state of the art transformer based summarization methods, but with 16 times fewer parameters. | Withdrawn |
ICLR.cc/2022/Conference | On Incorporating Inductive Biases into VAEs | We explain why directly changing the prior can be a surprisingly ineffective mechanism for incorporating inductive biases into variational auto-encoders (VAEs), and introduce a simple and effective alternative approach: Intermediary Latent Space VAEs (InteL-VAEs). InteL-VAEs use an intermediary set of latent variables to control the stochasticity of the encoding process, before mapping these in turn to the latent representation using a parametric function that encapsulates our desired inductive bias(es). This allows us to impose properties like sparsity or clustering on learned representations, and incorporate human knowledge into the generative model. Whereas changing the prior only indirectly encourages behavior through regularizing the encoder, InteL-VAEs are able to directly enforce desired characteristics. Moreover, they bypass the computation and encoder design issues caused by non-Gaussian priors, while allowing for additional flexibility through training of the parametric mapping function. We show that these advantages, in turn, lead to both better generative models and better representations being learned. | Accept (Poster) |
ICLR.cc/2023/Conference | What Knowledge gets Distilled in Knowledge Distillation? | Knowledge distillation aims to transfer useful information from a teacher network to a student network, with the primary goal of improving the student's performance for the task at hand. Over the years, there has a been a deluge of novel techniques and use cases of knowledge distillation. Yet, despite the various improvements, there seems to be a glaring gap in the community's fundamental understanding of the process. Specifically, what is the knowledge that gets distilled in knowledge distillation? In other words, in what ways does the student become similar to the teacher? Does it start to localize objects in the same way? Does it get fooled by the same adversarial samples? Does its data invariance properties become similar? Our work presents a comprehensive study to try to answer these questions and more. Our results, using image classification as a case study and three state-of-the-art knowledge distillation techniques, show that knowledge distillation methods can indeed indirectly distill other kinds of properties beyond improving task performance. By exploring these questions, we hope for our work to provide a clearer picture of what happens during knowledge distillation. | Withdrawn |
ICLR.cc/2019/Conference | Distribution-Interpolation Trade off in Generative Models | We investigate the properties of multidimensional probability distributions in the context of latent space prior distributions of implicit generative models. Our work revolves around the phenomena arising while decoding linear interpolations between two random latent vectors -- regions of latent space in close proximity to the origin of the space are oversampled, which restricts the usability of linear interpolations as a tool to analyse the latent space. We show that the distribution mismatch can be eliminated completely by a proper choice of the latent probability distribution or using non-linear interpolations. We prove that there is a trade off between the interpolation being linear, and the latent distribution having even the most basic properties required for stable training, such as finite mean. We use the multidimensional Cauchy distribution as an example of the prior distribution, and also provide a general method of creating non-linear interpolations, that is easily applicable to a large family of commonly used latent distributions. | Accept (Poster) |
ICLR.cc/2019/Conference | Minimal Images in Deep Neural Networks: Fragile Object Recognition in Natural Images | The human ability to recognize objects is impaired when the object is not shown in full. "Minimal images" are the smallest regions of an image that remain recognizable for humans. Ullman et al. (2016) show that a slight modification of the location and size of the visible region of the minimal image produces a sharp drop in human recognition accuracy. In this paper, we demonstrate that such drops in accuracy due to changes of the visible region are a common phenomenon between humans and existing state-of-the-art deep neural networks (DNNs), and are much more prominent in DNNs. We found many cases where DNNs classified one region correctly and the other incorrectly, though they only differed by one row or column of pixels, and were often bigger than the average human minimal image size. We show that this phenomenon is independent from previous works that have reported lack of invariance to minor modifications in object location in DNNs. Our results thus reveal a new failure mode of DNNs that also affects humans to a much lesser degree. They expose how fragile DNN recognition ability is in natural images even without adversarial patterns being introduced. Bringing the robustness of DNNs in natural images to the human level remains an open challenge for the community. | Accept (Poster) |
ICLR.cc/2021/Conference | VideoGen: Generative Modeling of Videos using VQ-VAE and Transformers | We present VideoGen: a conceptually simple architecture for scaling likelihood based generative modeling to natural videos. VideoGen uses VQ-VAE that learns learns downsampled discrete latent representations of a video by employing 3D convolutions and axial self-attention. A simple GPT-like architecture is then used to autoregressively model the discrete latents using spatio-temporal position encodings. Despite the simplicity in formulation, ease of training and a light compute requirement, our architecture is able to generate samples competitive with state-of-the-art GAN models for video generation on the BAIR Robot dataset, and generate coherent action-conditioned samples based on experiences gathered from the VizDoom simulator. We hope our proposed architecture serves as a reproducible reference for a minimalistic implementation of transformer based video generation models without requiring industry scale compute resources. Samples are available at https://sites.google.com/view/videogen | Reject |
ICLR.cc/2019/Conference | DEEP ADVERSARIAL FORWARD MODEL | Learning world dynamics has recently been investigated as a way to make reinforcement
learning (RL) algorithms to be more sample efficient and interpretable.
In this paper, we propose to capture an environment dynamics with a novel forward
model that leverages recent works on adversarial learning and visual control. Such
a model estimates future observations conditioned on the current ones and other
input variables such as actions taken by an RL-agent. We focus on image generation
which is a particularly challenging topic but our method can be adapted to
other modalities. More precisely, our forward model is trained to produce realistic
observations of the future while a discriminator model is trained to distinguish
between real images and the model’s prediction of the future. This approach overcomes
the need to define an explicit loss function for the forward model which is currently
used for solving such a class of problem. As a consequence, our learning protocol
does not have to rely on an explicit distance such as Euclidean distance which
tends to produce unsatisfactory predictions. To illustrate our method, empirical
qualitative and quantitative results are presented on a real driving scenario, along
with qualitative results on Atari game Frostbite. | Reject |
ICLR.cc/2022/Conference | Soteria: In search of efficient neural networks for private inference | In the context of ML as a service, our objective is to protect the confidentiality of the users’ queries and the server's model parameters, with modest computation and communication overhead. Prior solutions primarily propose fine-tuning cryptographic methods to make them efficient for known fixed model architectures. The drawback with this line of approach is that the model itself is never designed to efficiently operate with existing cryptographic computations. We observe that the network architecture, internal functions, and parameters of a model, which are all chosen during training, significantly influence the computation and communication overhead of a cryptographic method, during inference.Thus, we propose SOTERIA — a training method to construct model architectures that are by-design efficient for private inference. We use neural architecture search algorithms with the dual objective of optimizing the accuracy of the model and the overhead of using cryptographic primitives for secure inference. Given the flexibility of modifying a model during training, we find accurate models that are also efficient for private computation. We select garbled circuits as our underlying cryptographic primitive, due to their expressiveness and efficiency. We empirically evaluate SOTERIA on MNIST and CIFAR10 datasets, to compare with the prior work on secure inference. Our results confirm that SOTERIA is indeed effective in balancing performance and accuracy. | Reject |
ICLR.cc/2020/Conference | SVQN: Sequential Variational Soft Q-Learning Networks | Partially Observable Markov Decision Processes (POMDPs) are popular and flexible models for real-world decision-making applications that demand the information from past observations to make optimal decisions. Standard reinforcement learning algorithms for solving Markov Decision Processes (MDP) tasks are not applicable, as they cannot infer the unobserved states. In this paper, we propose a novel algorithm for POMDPs, named sequential variational soft Q-learning networks (SVQNs), which formalizes the inference of hidden states and maximum entropy reinforcement learning (MERL) under a unified graphical model and optimizes the two modules jointly. We further design a deep recurrent neural network to reduce the computational complexity of the algorithm. Experimental results show that SVQNs can utilize past information to help decision making for efficient inference, and outperforms other baselines on several challenging tasks. Our ablation study shows that SVQNs have the generalization ability over time and are robust to the disturbance of the observation. | Accept (Poster) |
ICLR.cc/2023/Conference | Rethinking Learning Dynamics in RL using Adversarial Networks | Recent years have seen tremendous progress in methods of reinforcement learning. However, most of these approaches have been trained in a straightforward fashion and are generally not robust to adversity, especially in the meta-RL setting. To the best of our knowledge, our work is the first to propose an adversarial training regime for Multi-Task Reinforcement Learning, which requires no manual intervention or domain knowledge of the environments. Our experiments on multiple environments in the Multi-Task Reinforcement learning domain demonstrate that the adversarial process leads to a better exploration of numerous solutions and a deeper understanding of the environment. We also adapt existing measures of causal attribution to draw insights from the skills learned, facilitating easier re-purposing of skills for adaptation to unseen environments and tasks.
| Withdrawn |
ICLR.cc/2021/Conference | Neural CDEs for Long Time Series via the Log-ODE Method | Neural Controlled Differential Equations (Neural CDEs) are the continuous-time analogue of an RNN, just as Neural ODEs are analogous to ResNets. However just like RNNs, training Neural CDEs can be difficult for long time series. Here, we propose to apply a technique drawn from stochastic analysis, namely the log-ODE method. Instead of using the original input sequence, our procedure summarises the information over local time intervals via the log-signature map, and uses the resulting shorter stream of log-signatures as the new input. This represents a length/channel trade-off. In doing so we demonstrate efficacy on problems of length up to 17k observations and observe significant training speed-ups, improvements in model performance, and reduced memory requirements compared to the existing algorithm. | Reject |
ICLR.cc/2020/Conference | R2D2: Reuse & Reduce via Dynamic Weight Diffusion for Training Efficient NLP Models | We propose R2D2 layers, a new neural block for training efficient NLP models. Our proposed method is characterized by a dynamic weight diffusion mechanism which learns to reuse and reduce parameters in the conventional transformation layer, commonly found in popular Transformer/LSTMs models. Our method is inspired by recent Quaternion methods which share parameters via the Hamilton product. This can be interpreted as a neural and learned approximation of the Hamilton product which imbues our method with increased flexibility and expressiveness, i.e., we are no longer restricted by the 4D nature of Quaternion weight sharing. We conduct extensive experiments in the NLP domain, showing that R2D2 (i) enables a parameter savings of up to 2 times to 16 times with minimal degradation of performance and (ii) outperforms other parameter savings alternative such as low-rank factorization and Quaternion methods.
| Reject |
ICLR.cc/2020/Conference | Accelerating Reinforcement Learning Through GPU Atari Emulation | We introduce CuLE (CUDA Learning Environment), a CUDA port of the Atari Learning Environment (ALE) which is used for the development of deep reinforcement algorithms. CuLE overcomes many limitations of existing CPU-based emulators and scales naturally to multiple GPUs. It leverages GPU parallelization to run thousands of games simultaneously and it renders frames directly on the GPU, to avoid the bottleneck arising from the limited CPU-GPU communication bandwidth. CuLE generates up to 155M frames per hour on a single GPU, a finding previously achieved only through a cluster of CPUs. Beyond highlighting the differences between CPU and GPU emulators in the context of reinforcement learning, we show how to leverage the high throughput of CuLE by effective batching of the training data, and show accelerated convergence for A2C+V-trace. CuLE is available at [hidden URL]. | Reject |
ICLR.cc/2023/Conference | Go-Explore with a guide: Speeding up search in sparse reward settings with goal-directed intrinsic rewards | Reinforcement Learning (RL) agents have traditionally been very sample-intensive to train, especially in environments with sparse rewards. Seeking inspiration from neuroscience experiments of rats learning the structure of a maze without needing extrinsic rewards, we seek to incorporate additional intrinsic rewards to guide behavior. We propose a potential-based goal-directed intrinsic reward (GDIR), which provides a reward signal regardless of whether the task is achieved, and ensures that learning can always take place. While GDIR may be similar to approaches such as reward shaping in incorporating goal-based rewards, we highlight that GDIR is innate to the agent and hence applicable across a wide range of environments without needing to rely on a properly shaped environment reward. We also note that GDIR is different from curiosity-based intrinsic motivation, which can diminish over time and lead to inefficient exploration. Go-Explore is a well-known state-of-the-art algorithm for sparse reward domains, and we demonstrate that by incorporating GDIR in the ``Go" and ``Explore" phases, we can improve Go-Explore's performance and enable it to learn faster across multiple environments, for both discrete (2D grid maze environments, Towers of Hanoi, Game of Nim) and continuous (Cart Pole and Mountain Car) state spaces. Furthermore, to consolidate learnt trajectories better, our method also incorporates a novel approach of hippocampal replay to update the values of GDIR and reset state visit and selection counts of states along the successful trajectory. As a benchmark, we also show that our proposed approaches learn significantly faster than traditional extrinsic-reward-based RL algorithms such as Proximal Policy Optimization, TD-learning, and Q-learning. | Reject |
ICLR.cc/2019/Conference | Neural Probabilistic Motor Primitives for Humanoid Control | We focus on the problem of learning a single motor module that can flexibly express a range of behaviors for the control of high-dimensional physically simulated humanoids. To do this, we propose a motor architecture that has the general structure of an inverse model with a latent-variable bottleneck. We show that it is possible to train this model entirely offline to compress thousands of expert policies and learn a motor primitive embedding space. The trained neural probabilistic motor primitive system can perform one-shot imitation of whole-body humanoid behaviors, robustly mimicking unseen trajectories. Additionally, we demonstrate that it is also straightforward to train controllers to reuse the learned motor primitive space to solve tasks, and the resulting movements are relatively naturalistic. To support the training of our model, we compare two approaches for offline policy cloning, including an experience efficient method which we call linear feedback policy cloning. We encourage readers to view a supplementary video (https://youtu.be/CaDEf-QcKwA ) summarizing our results. | Accept (Poster) |
ICLR.cc/2022/Conference | Geometric Algebra Attention Networks for Small Point Clouds | Much of the success of deep learning is drawn from building architectures that properly respect underlying symmetry and structure in the data on which they operate—a set of considerations that have been united under the banner of geometric deep learning. Often problems in the physical sciences deal with relatively small sets of points in two- or three-dimensional space wherein translation, rotation, and permutation equivariance are important or even vital for models to be useful in practice. In this work, we present rotation- and permutation-equivariant architectures for deep learning on these small point clouds, composed of a set of products of terms from the geometric algebra and reductions over those products using an attention mechanism. The geometric algebra provides valuable mathematical structure by which to combine vector, scalar, and other types of geometric inputs in a systematic way to account for rotation invariance or covariance, while attention yields a powerful way to impose permutation equivariance. We demonstrate the usefulness of these architectures by training models to solve sample problems relevant to physics, chemistry, and biology.
| Reject |
ICLR.cc/2021/Conference | Reinforcement Learning for Sparse-Reward Object-Interaction Tasks in First-person Simulated 3D Environments | First-person object-interaction tasks in high-fidelity, 3D, simulated environments such as the AI2Thor virtual home-environment pose significant sample-efficiency challenges for reinforcement learning (RL) agents learning from sparse task rewards. To alleviate these challenges, prior work has provided extensive supervision via a combination of reward-shaping, ground-truth object-information, and expert demonstrations. In this work, we show that one can learn object-interaction tasks from scratch without supervision by learning an attentive object-model as an auxiliary task during task learning with an object-centric relational RL agent. Our key insight is that learning an object-model that incorporates object-relationships into forward prediction provides a dense learning signal for unsupervised representation learning of both objects and their relationships. This, in turn, enables faster policy learning for an object-centric relational RL agent. We demonstrate our agent by introducing a set of challenging object-interaction tasks in the AI2Thor environment where learning with our attentive object-model is key to strong performance. Specifically, by comparing our agent and relational RL agents with alternative auxiliary tasks with a relational RL agent equipped with ground-truth object-information, we find that learning with our object-model best closes the performance gap in terms of both learning speed and maximum success rate. Additionally, we find that incorporating object-relationships into an object-model's forward predictions is key to learning representations that capture object-category and object-state. | Withdrawn |
ICLR.cc/2020/Conference | Improving Sequential Latent Variable Models with Autoregressive Flows | We propose an approach for sequence modeling based on autoregressive normalizing flows. Each autoregressive transform, acting across time, serves as a moving reference frame for modeling higher-level dynamics. This technique provides a simple, general-purpose method for improving sequence modeling, with connections to existing and classical techniques. We demonstrate the proposed approach both with standalone models, as well as a part of larger sequential latent variable models. Results are presented on three benchmark video datasets, where flow-based dynamics improve log-likelihood performance over baseline models. | Reject |
ICLR.cc/2022/Conference | Factored World Models for Zero-Shot Generalization in Robotic Manipulation | World models for environments with many objects face a combinatorial explosion of states: as the number of objects increases, the number of possible arrangements grows exponentially. In this paper, we learn to generalize over robotic pick-and-place tasks using object-factored world models, which combat the combinatorial explosion by ensuring that predictions are equivariant to permutations of objects. We build on one such model, C-SWM, which we extend to overcome the assumption that each action is associated with one object. To do so, we introduce an action attention module to determine which objects are likely to be affected by an action. The attention module is used in conjunction with a residual graph neural network block that receives action information at multiple levels. Based on RGB images and parameterized motion primitives, our model can accurately predict the dynamics of a robot building structures from blocks of various shapes. Our model generalizes over training structures built in different positions. Moreover crucially, the learned model can make predictions about tasks not represented in training data. That is, we demonstrate successful zero-shot generalization to novel tasks. For example, we measure only 2.4% absolute decrease in our action ranking metric in the case of a block assembly task. | Reject |
ICLR.cc/2021/Conference | Single Pair Cross-Modality Super Resolution | Non-visual imaging sensors are widely used in the industry for different purposes. Those sensors are more expensive than visual (RGB) sensors, and usually produce images with lower resolution. To this end, Cross-Modality Super-Resolution methods were introduced, where an RGB image of a high-resolution assists in increasing the resolution of a low-resolution modality. However, fusing images from different modalities is not a trivial task, since each multi-modal pair varies greatly in its internal correlations. For this reason, traditional state-of-the-arts which are trained on external datasets often struggle with yielding an artifact-free result that is still loyal to the target modality characteristics.
We present CMSR, a single-pair approach for Cross-Modality Super-Resolution. The network is internally trained on the two input images only, in a self-supervised manner, learns their internal statistics and correlations, and applies them to upsample the target modality. CMSR contains an internal transformer which is trained on-the-fly together with the up-sampling process itself and without supervision, to allow dealing with pairs that are only weakly aligned. We show that CMSR produces state-of-the-art super resolved images, yet without introducing artifacts or irrelevant details that originate from the RGB image only. | Withdrawn |
ICLR.cc/2019/Conference | DyRep: Learning Representations over Dynamic Graphs | Representation Learning over graph structured data has received significant attention recently due to its ubiquitous applicability. However, most advancements have been made in static graph settings while efforts for jointly learning dynamic of the graph and dynamic on the graph are still in an infant stage. Two fundamental questions arise in learning over dynamic graphs: (i) How to elegantly model dynamical processes over graphs? (ii) How to leverage such a model to effectively encode evolving graph information into low-dimensional representations? We present DyRep - a novel modeling framework for dynamic graphs that posits representation learning as a latent mediation process bridging two observed processes namely -- dynamics of the network (realized as topological evolution) and dynamics on the network (realized as activities between nodes). Concretely, we propose a two-time scale deep temporal point process model that captures the interleaved dynamics of the observed processes. This model is further parameterized by a temporal-attentive representation network that encodes temporally evolving structural information into node representations which in turn drives the nonlinear evolution of the observed graph dynamics. Our unified framework is trained using an efficient unsupervised procedure and has capability to generalize over unseen nodes. We demonstrate that DyRep outperforms state-of-the-art baselines for dynamic link prediction and time prediction tasks and present extensive qualitative insights into our framework. | Accept (Poster) |
ICLR.cc/2022/Conference | The Uncanny Similarity of Recurrence and Depth | It is widely believed that deep neural networks contain layer specialization, wherein networks extract hierarchical features representing edges and patterns in shallow layers and complete objects in deeper layers. Unlike common feed-forward models that have distinct filters at each layer, recurrent networks reuse the same parameters at various depths. In this work, we observe that recurrent models exhibit the same hierarchical behaviors and the same performance benefits as depth despite reusing the same filters at every recurrence. By training models of various feed-forward and recurrent architectures on several datasets for image classification as well as maze solving, we show that recurrent networks have the ability to closely emulate the behavior of non-recurrent deep models, often doing so with far fewer parameters. | Accept (Poster) |
ICLR.cc/2022/Conference | Sharpness-Aware Minimization in Large-Batch Training: Training Vision Transformer In Minutes | Large-batch training is an important direction for distributed machine learning, which can improve the utilization of large-scale clusters and therefore accelerate the training process. However, recent work illustrates that large-batch training is prone to converge to sharp minima and cause a huge generalization gap. Sharpness-Aware Minimization (SAM) tries to narrow the generalization gap by seeking parameters that lie in a flat region. However, it requires two sequential gradient calculations that doubles the computational overhead. In this paper, we propose a novel algorithm LookSAM to significantly reduce its additional training cost. We further propose a layer-wise modification for adapting LookSAM to the large-batch training setting (Look-LayerSAM). Equipped with our enhanced training algorithm, we are the first to successfully scale up the batch size when training Vision Transformers (ViTs). With a 64k batch size, we are able to train ViTs from scratch within an hour while maintaining competitive performance. | Withdrawn |
ICLR.cc/2018/Conference | Statestream: A toolbox to explore layerwise-parallel deep neural networks | Building deep neural networks to control autonomous agents which have to interact in real-time with the physical world, such as robots or automotive vehicles, requires a seamless integration of time into a network’s architecture. The central question of this work is, how the temporal nature of reality should be reflected in the execution of a deep neural network and its components. Most artificial deep neural networks are partitioned into a directed graph of connected modules or layers and the layers themselves consist of elemental building blocks, such as single units. For most deep neural networks, all units of a layer are processed synchronously and in parallel, but layers themselves are processed in a sequential manner. In contrast, all elements of a biological neural network are processed in parallel. In this paper, we define a class of networks between these two extreme cases. These networks are executed in a streaming or synchronous layerwise-parallel manner, unlocking the layers of such networks for parallel processing. Compared to the standard layerwise-sequential deep networks, these new layerwise-parallel networks show a fundamentally different temporal behavior and flow of information, especially for networks with skip or recurrent connections. We argue that layerwise-parallel deep networks are better suited for future challenges of deep neural network design, such as large functional modularized and/or recurrent architectures as well as networks allocating different network capacities dependent on current stimulus and/or task complexity. We layout basic properties and discuss major challenges for layerwise-parallel networks. Additionally, we provide a toolbox to design, train, evaluate, and online-interact with layerwise-parallel networks. | Reject |
ICLR.cc/2021/Conference | Language-Agnostic Representation Learning of Source Code from Structure and Context | Source code (Context) and its parsed abstract syntax tree (AST; Structure) are two complementary representations of the same computer program. Traditionally, designers of machine learning models have relied predominantly either on Structure or Context. We propose a new model, which jointly learns on Context and Structure of source code. In contrast to previous approaches, our model uses only language-agnostic features, i.e., source code and features that can be computed directly from the AST. Besides obtaining state-of-the-art on monolingual code summarization on all five programming languages considered in this work, we propose the first multilingual code summarization model. We show that jointly training on non-parallel data from multiple programming languages improves results on all individual languages, where the strongest gains are on low-resource languages. Remarkably, multilingual training only from Context does not lead to the same improvements, highlighting the benefits of combining Structure and Context for representation learning on code. | Accept (Poster) |
ICLR.cc/2022/Conference | Towards the Memorization Effect of Neural Networks in Adversarial Training | Recent studies suggest that "memorization" is one important factor for overparameterized deep neural networks (DNNs) to achieve optimal performance. Specifically, the perfectly fitted DNNs can memorize the labels of many atypical samples, generalize their memorization to correctly classify test atypical samples and enjoy better test performance. While, DNNs which are optimized via adversarial training algorithms can also achieve perfect training performance by memorizing the labels of atypical samples, as well as the adversarially perturbed atypical samples. However, adversarially trained models always suffer from poor generalization, with both relatively low clean accuracy and robustness on the test set. In this work, we study the effect of memorization in adversarial trained DNNs and disclose two important findings: (a) Memorizing atypical samples is only effective to improve DNN's accuracy on clean atypical samples, but hardly improve their adversarial robustness and (b) Memorizing certain atypical samples will even hurt the DNN's performance on typical samples. Based on these two findings, we propose Benign Adversarial Training (BAT) which can facilitate adversarial training to avoid fitting "harmful" atypical samples and fit as more "benign" atypical samples as possible. In our experiments, we validate the effectiveness of BAT, and show it can achieve better clean accuracy vs. robustness trade-off than baseline methods, in benchmark datasets such as CIFAR100 and Tiny ImageNet. | Reject |
ICLR.cc/2023/Conference | Short-Term Memory Convolutions | The real-time processing of time series signals is a critical issue for many real-life applications. The idea of real-time processing is especially important in audio domain as the human perception of sound is sensitive to any kind of disturbance in perceived signals, especially the lag between auditory and visual modalities. The rise of deep learning (DL) models complicated the landscape of signal processing. Although they often have superior quality compared to standard DSP methods, this advantage is diminished by higher latency. In this work we propose novel method for minimization of inference time latency and memory consumption, called Short-Term Memory Convolution (STMC) and its transposed counterpart. The main advantage of STMC is the low latency comparable to long short-term memory (LSTM) networks. Furthermore, the training of STMC-based models is faster and more stable as the method is based solely on convolutional neural networks (CNNs). In this study we demonstrate an application of this solution to a U-Net model for a speech separation task and GhostNet model in acoustic scene classification (ASC) task. In case of speech separation we achieved a 5-fold reduction in inference time and a 2-fold reduction in latency without affecting the output quality. The inference time for ASC task was up to 4 times faster while preserving the original accuracy. | Accept: poster |
ICLR.cc/2023/Conference | Graph Backup: Data Efficient Backup Exploiting Markovian Transitions | The successes of deep Reinforcement Learning (RL) are limited to settings where we have a large stream of online experiences, but applying RL in the data-efficient setting with limited access to online interactions is still challenging. A key to data-efficient RL is good value estimation, but current methods in this space fail to fully utilise the structure of the trajectory data gathered from the environment. In this paper, we treat the transition data of the MDP as a graph, and define a novel backup operator, Graph Backup, which exploits this graph structure for better value estimation. Compared to multi-step backup methods such as $n$-step $Q$-Learning and TD($\lambda$), Graph Backup can perform counterfactual credit assignment and gives stable value estimates for a state regardless of which trajectory the state is sampled from. Our method, when combined with popular off-policy value-based methods, provides improved performance over one-step and multi-step methods on a suite of data-efficient RL benchmarks including MiniGrid, Minatar and Atari100K. We further analyse the reasons for this performance boost through a novel visualisation of the transition graphs of Atari games. | Reject |
ICLR.cc/2023/Conference | Generating Adversarial Examples with Task Oriented Multi-Objective Optimization | Deep learning models, even the-state-of-the-art ones, are highly vulnerable to adversarial examples. Adversarial training is one of the most efficient methods to improve the model's robustness. The key factor for the success of adversarial training is the capability to generate qualified and divergent adversarial examples which satisfy some objectives/goals (e.g., finding adversarial examples that maximize the model losses for simultaneously attacking multiple models). Therefore, multi-objective optimization (MOO) is a natural tool for adversarial example generation, where we search adversarial examples simultaneously maximizing some objectives/goals. However, we observe that a naive application of MOO tends to maximize all objectives/goals equally, without caring if an objective/goal has been achieved yet. This leads to useless effort to further improve the goal-achieved tasks, while putting less focus on the goal-unachieved tasks. In this paper, we propose \emph{Task Oriented MOO} to address this issue, in the context where we can explicitly define the goal achievement for a task. Our principle is to only maintain the goal-achieved tasks, while letting the optimizer spend more effort on improving the goal-unachieved tasks. We conduct comprehensive experiments for our Task Oriented MOO on various adversarial example generation schemes. The experimental results firmly demonstrate the merit of our proposed approach. | Reject |
ICLR.cc/2023/Conference | DeepTime: Deep Time-index Meta-learning for Non-stationary Time-series Forecasting | Advances in I.T. infrastructure has led to the collection of longer sequences of time-series. Such sequences are typically non-stationary, exhibiting distribution shifts over time -- a challenging scenario for the forecasting task, due to the problems of covariate shift, and conditional distribution shift. In this paper, we show that deep time-index models possess strong synergies with a meta-learning formulation of forecasting, displaying significant advantages over existing neural forecasting methods in tackling the problems arising from non-stationarity. These advantages include having a stronger smoothness prior, avoiding the problem of covariate shift, and having better sample efficiency. To this end, we propose DeepTime, a deep time-index model trained via meta-learning. Extensive experiments on real-world datasets in the long sequence time-series forecasting setting demonstrate that our approach achieves competitive results with state-of-the-art methods, and is highly efficient. Code is attached as supplementary material, and will be publicly released. | Reject |
ICLR.cc/2023/Conference | Improve the Adaptation Process by Reasoning From Failed and Successful Cases | Usually, existing works on adaptation in reasoning-based systems assume that the case base holds only successful cases, i.e., cases having solutions believed to be appropriate for the corresponding problems. However, in practice, the case base could hold failed cases, resulting from an earlier adaptation process but discarded by the revision process. Not considering failed cases would be missing an interesting opportunity to learn more knowledge for improving the adaptation process.
This paper proposes a novel approach to the adaptation process in the case-based reasoning paradigm, based on an improved barycentric approach by considering the failed cases.The experiment performed on real data demonstrates the benefit of the method considering the failed cases in the adaptation process compared to the classical ones that ignore them, thus, improving the performance of the case-based reasoning system. | Withdrawn |
ICLR.cc/2020/Conference | Generalized Transformation-based Gradient | The reparameterization trick has become one of the most useful tools in the field of variational inference. However, the reparameterization trick is based on the standardization transformation which restricts the scope of application of this method to distributions that have tractable inverse cumulative distribution functions or are expressible as deterministic transformations of such distributions. In this paper, we generalized the reparameterization trick by allowing a general transformation. Unlike other similar works, we develop the generalized transformation-based gradient model formally and rigorously. We discover that the proposed model is a special case of control variate indicating that the proposed model can combine the advantages of CV and generalized reparameterization. Based on the proposed gradient model, we propose a new polynomial-based gradient estimator which has better theoretical performance than the reparameterization trick under certain condition and can be applied to a larger class of variational distributions. In studies of synthetic and real data, we show that our proposed gradient estimator has a significantly lower gradient variance than other state-of-the-art methods thus enabling a faster inference procedure. | Withdrawn |
ICLR.cc/2020/Conference | Efficient Riemannian Optimization on the Stiefel Manifold via the Cayley Transform | Strictly enforcing orthonormality constraints on parameter matrices has been shown advantageous in deep learning. This amounts to Riemannian optimization on the Stiefel manifold, which, however, is computationally expensive. To address this challenge, we present two main contributions: (1) A new efficient retraction map based on an iterative Cayley transform for optimization updates, and (2) An implicit vector transport mechanism based on the combination of a projection of the momentum and the Cayley transform on the Stiefel manifold. We specify two new optimization algorithms: Cayley SGD with momentum, and Cayley ADAM on the Stiefel manifold. Convergence of Cayley SGD is theoretically analyzed. Our experiments for CNN training demonstrate that both algorithms: (a) Use less running time per iteration relative to existing approaches that enforce orthonormality of CNN parameters; and (b) Achieve faster convergence rates than the baseline SGD and ADAM algorithms without compromising the performance of the CNN. Cayley SGD and Cayley ADAM are also shown to reduce the training time for optimizing the unitary transition matrices in RNNs. | Accept (Poster) |
ICLR.cc/2020/Conference | Learning to Rank Learning Curves | Many automated machine learning methods, such as those for hyperparameter and neural architecture optimization, are computationally expensive because they involve training many different model configurations. In this work, we present a new method that saves computational budget by terminating poor configurations early on in the training. In contrast to existing methods, we consider this task as a ranking and transfer learning problem. We qualitatively show that by optimizing a pairwise ranking loss and leveraging learning curves from other data sets, our model is able to effectively rank learning curves without having to observe many or very long learning curves. We further demonstrate that our method can be used to accelerate a neural architecture search by a factor of up to 100 without a significant performance degradation of the discovered architecture. In further experiments we analyze the quality of ranking, the influence of different model components as well as the predictive behavior of the model. | Reject |
ICLR.cc/2020/Conference | Meta Module Network for Compositional Visual Reasoning | There are two main lines of research on visual reasoning: neural module network (NMN) with explicit multi-hop reasoning through handcrafted neural modules, and monolithic network with implicit reasoning in the latent feature space. The former excels in interpretability and compositionality, while the latter usually achieves better performance due to model flexibility and parameter efficiency.
In order to bridge the gap of the two, we present Meta Module Network (MMN), a novel hybrid approach that can efficiently utilize a Meta Module to perform versatile functionalities, while preserving compositionality and interpretability through modularized design. The proposed model first parses an input question into a functional program through a Program Generator. Instead of handcrafting a task-specific network to represent each function like traditional NMN, we use Recipe Encoder to translate the functions into their corresponding recipes (specifications), which are used to dynamically instantiate the Meta Module into Instance Modules. To endow different instance modules with designated functionality, a Teacher-Student framework is proposed, where a symbolic teacher pre-executes against the scene graphs to provide guidelines for the instantiated modules (student) to follow. In a nutshell, MMN adopts the meta module to increase its parameterization efficiency, and uses recipe encoding to improve its generalization ability over NMN. Experiments conducted on the GQA benchmark demonstrates that: (1) MMN achieves significant improvement over both NMN and monolithic network baselines; (2) MMN is able to generalize to unseen but related functions. | Withdrawn |
ICLR.cc/2023/Conference | Language Model Pre-training with Linguistically Motivated Curriculum Learning | Pre-training serves as a foundation of recent NLP models, where language modeling task is performed over large texts. It has been shown that data affects the quality of pre-training, and curriculum has been investigated regarding sequence length. We consider a linguistic perspective in the curriculum, where frequent words are learned first and rare words last. This is achieved by substituting syntactic constituents for rare words with their constituent labels. By such syntactic substitutions, a curriculum can be made by gradually introducing words with decreasing frequency levels. Without modifying model architectures or introducing external computational overhead, our data-centric method gives better performances over vanilla BERT on various downstream benchmarks. | Reject |
ICLR.cc/2023/Conference | Parallel Federated Learning over Heterogeneous Devices | Federated Learning (FL) is a cutting-edge distributed machine learning framework that enables multiple devices to collaboratively train a shared model without exposing their own data. In the scenario of device heterogeneity, the synchronous FL suffers from latency bottleneck induced by network stragglers, which hampers the training efficiency significantly. In addition, due to the diverse structures and sizes of local models, the simple and fast averaging aggregation is not feasible anymore. Instead, complicated aggregation operation, such as knowledge distillation, is required. The time cost for complicated aggregation becomes a new bottleneck that limits the computational efficiency of FL.
In this work, we claim that the cause root of training latency actually lies in the aggregation-then-broadcasting workflow of the server. By swapping the computational order of aggregation and broadcasting, we propose a new parallel federated learning (PFL) framework, which unlocks the edge nodes during global computation and the central server during local computation. This fully asynchronous and parallel pipeline enables handling device heterogeneity and network stragglers, allowing flexible device participation as well as achieving scalability in computation.
We theoretically prove that PFL can achieve the similar convergence rate as synchronous FL, and empirically show that our framework can tolerate both stragglers and complicated aggregation tasks, which brings $1.77\times$ to $7.32\times$ speedup. | Withdrawn |
ICLR.cc/2020/Conference | Ensemble methods and LSTM outperformed other eight machine learning classifiers in an EEG-based BCI experiment | We review eight machine learning classification algorithms to analyze Electroencephalographic (EEG) signals in order to distinguish EEG patterns associated with five basic educational tasks. There is a large variety of classifiers being used in this EEG-based Brain-Computer Interface (BCI) field. While previous EEG experiments used several classifiers in the same experiments or reviewed different algorithms on datasets from different experiments, our approach focuses on review eight classifier categories on the same dataset, including linear classifiers, non-linear Bayesian classifiers, nearest neighbour classifiers, ensemble methods, adaptive classifiers, tensor classifiers, transfer learning and deep learning. Besides, we intend to find an approach which can run smoothly on the current mainstream personal computers and smartphones. The empirical evaluation demonstrated that Random Forest and LSTM (Long Short-Term Memory) outperform other approaches. We used a data set which users were conducting five frequently-conduct learning-related tasks, including reading, writing, and typing. Results showed that these best two algorithms could correctly classify different users with an accuracy increase of 5% to 9%, use each task independently. Within each subject, the tasks could be recognized with an accuracy increase of 4% to 7%, compared with other approaches. This work suggests that Random Forest could be a recommended approach (fast and accurate) for current mainstream hardware, while LSTM has the potential to be the first-choice approach when the mainstream computers and smartphones can process more data in a shorter time. | Desk_Rejected |
ICLR.cc/2018/Conference | Entropy-SGD optimizes the prior of a PAC-Bayes bound: Data-dependent PAC-Bayes priors via differential privacy | We show that Entropy-SGD (Chaudhari et al., 2017), when viewed as a learning algorithm, optimizes a PAC-Bayes bound on the risk of a Gibbs (posterior) classifier, i.e., a randomized classifier obtained by a risk-sensitive perturbation of the weights of a learned classifier. Entropy-SGD works by optimizing the bound’s prior, violating the hypothesis of the PAC-Bayes theorem that the prior is chosen independently of the data. Indeed, available implementations of Entropy-SGD rapidly obtain zero training error on random labels and the same holds of the Gibbs posterior. In order to obtain a valid generalization bound, we show that an ε-differentially private prior yields a valid PAC-Bayes bound, a straightforward consequence of results connecting generalization with differential privacy. Using stochastic gradient Langevin dynamics (SGLD) to approximate the well-known exponential release mechanism, we observe that generalization error on MNIST (measured on held out data) falls within the (empirically nonvacuous) bounds computed under the assumption that SGLD produces perfect samples. In particular, Entropy-SGLD can be configured to yield relatively tight generalization bounds and still fit real labels, although these same settings do not obtain state-of-the-art performance. | Reject |
ICLR.cc/2020/Conference | Zeno++: Robust Fully Asynchronous SGD | We propose Zeno++, a new robust asynchronous Stochastic Gradient Descent~(SGD) procedure which tolerates Byzantine failures of the workers. In contrast to previous work, Zeno++ removes some unrealistic restrictions on worker-server communications, allowing for fully asynchronous updates from anonymous workers, arbitrarily stale worker updates, and the possibility of an unbounded number of Byzantine workers. The key idea is to estimate the descent of the loss value after the candidate gradient is applied, where large descent values indicate that the update results in optimization progress. We prove the convergence of Zeno++ for non-convex problems under Byzantine failures. Experimental results show that Zeno++ outperforms existing approaches. | Reject |
ICLR.cc/2023/Conference | When and Why Vision-Language Models Behave like Bags-Of-Words, and What to Do About It? | Despite the success of large vision and language models (VLMs) in many downstream applications, it is unclear how well they encode the compositional relationships between objects and attributes. Here, we create the Attribution, Relation, and Order (ARO) benchmark to systematically evaluate the ability of VLMs to understand different types of relationships, attributes, and order information. ARO consists of \emph{Visual Genome Attribution}, to test the understanding of objects' properties; \emph{Visual Genome Relation}, to test for relational understanding; and \emph{COCO-Order \& Flickr30k-Order}, to test for order sensitivity in VLMs. ARO is orders of magnitude larger than previous benchmarks of compositionality, with more than 50,000 test cases. We present the settings where state-of-the-art VLMs behave like bags-of-words---i.e. when they have poor relational understanding, can blunder when linking objects to their attributes, and demonstrate a severe lack of order sensitivity. VLMs are predominantly trained and evaluated on large scale datasets with rich compositional structure in the images and captions. Yet, training on these datasets has not been enough to address the lack of compositional understanding, and evaluating on these datasets has failed to surface this deficiency. To understand why these limitations emerge and are not represented in the standard tests, we zoom into the evaluation and training procedures. We demonstrate that it is possible to perform well on image-text retrieval over existing datasets without using the composition and order information. This further motivates the value of using ARO to benchmark VLMs. Given that contrastive pretraining optimizes for retrieval on large datasets with similar shortcuts, we hypothesize that this can explain why the models do not need to learn to represent compositional information. This finding suggests a natural solution: composition-aware hard negative mining. We show that a simple-to-implement modification of contrastive learning significantly improves the performance on tasks requiring understanding of order and compositionality. | Accept: notable-top-5% |
ICLR.cc/2022/Conference | Boosting Randomized Smoothing with Variance Reduced Classifiers | Randomized Smoothing (RS) is a promising method for obtaining robustness certificates by evaluating a base model under noise. In this work, we: (i) theoretically motivate why ensembles are a particularly suitable choice as base models for RS, and (ii) empirically confirm this choice, obtaining state-of-the-art results in multiple settings. The key insight of our work is that the reduced variance of ensembles over the perturbations introduced in RS leads to significantly more consistent classifications for a given input. This, in turn, leads to substantially increased certifiable radii for samples close to the decision boundary. Additionally, we introduce key optimizations which enable an up to 55-fold decrease in sample complexity of RS for predetermined radii, thus drastically reducing its computational overhead. Experimentally, we show that ensembles of only 3 to 10 classifiers consistently improve on their strongest constituting model with respect to their average certified radius (ACR) by 5% to 21% on both CIFAR10 and ImageNet, achieving a new state-of-the-art ACR of 0.86 and 1.11, respectively. We release all code and models required to reproduce our results at https://github.com/eth-sri/smoothing-ensembles. | Accept (Spotlight) |
ICLR.cc/2021/Conference | Is Attention Better Than Matrix Decomposition? | As an essential ingredient of modern deep learning, attention mechanism, especially self-attention, plays a vital role in the global correlation discovery. However, is hand-crafted attention irreplaceable when modeling the global context? Our intriguing finding is that self-attention is not better than the matrix decomposition~(MD) model developed 20 years ago regarding the performance and computational cost for encoding the long-distance dependencies. We model the global context issue as a low-rank completion problem and show that its optimization algorithms can help design global information blocks. This paper then proposes a series of Hamburgers, in which we employ the optimization algorithms for solving MDs to factorize the input representations into sub-matrices and reconstruct a low-rank embedding. Hamburgers with different MDs can perform favorably against the popular global context module self-attention when carefully coping with gradients back-propagated through MDs. Comprehensive experiments are conducted in the vision tasks where it is crucial to learn the global context, including semantic segmentation and image generation, demonstrating significant improvements over self-attention and its variants. Code is available at https://github.com/Gsunshine/Enjoy-Hamburger. | Accept (Poster) |
ICLR.cc/2023/Conference | Sensitivity-aware Visual Parameter-efficient Tuning | Visual Parameter-efficient Tuning (VPT) has become a powerful alternative for full fine-tuning, which only updates a small number of parameters while freezing the remaining vast majority of parameters to significantly reduce the storage costs for adapting the pre-trained vision models to downstream tasks. Although the storage burden is largely alleviated, VPT approaches still face many challenges, e.g., lower inference speed and lacking effective configurations for trainable parameters tailored for each task. In this paper, we present a simple yet effective approach termed Sensitivity-aware visual Parameter-efficient Tuning (SPT) to tackle these challenges. Given a desired tunable parameter budget, SPT quickly identifies the important parameters to the given task in a data-dependent way before fine-tuning, without the complex selection schedule. To increase the representational capacity at a negligible cost within the same parameter budget, we employ low-rank reparameterization to achieve a better trade-off between parameter efficiency and accuracy. Through extensive experiments on a wide range of downstream recognition tasks, our SPT achieves better overall transfer performance than the full fine-tuning and the other VPT approaches, with no additional computational or memory overhead during inference. For instance, SPT saves 99.35% of the trainable parameters than the full fine-tuning while achieving a 7.3% higher average top-1 accuracy on VTAB-1k benchmark with the supervised pre-trained ViT-B backbone. Notably, SPT is also the first work that bridges the gap between full fine-tuning and VPT approaches for backbones under self-supervised pre-training strategies MAE and MoCo v3 on the challenging VTAB-1k benchmark. | Reject |
ICLR.cc/2023/Conference | Learning with Non-Uniform Label Noise: A Cluster-Dependent Semi-Supervised Approach | Learning with noisy labels is a challenging task in machine learning. Most existing methods explicitly or implicitly assume uniform label noise across all samples. In reality, label noise can be highly non-uniform in the feature space, with higher error rate for more difficult samples. Some recent works consider instance-dependent label noise but they require additional information such as some cleanly labeled data and confidence scores, which are usually unavailable or costly to obtain. In this paper, we consider learning with non-uniform label noise that requires no such additional information. we propose a cluster-dependent sample selection algorithm followed by a semi-supervised training mechanism based on the cluster-dependent label noise. The proposed self-adaptive multi-scale sample selection method increases the consistency of sample space by forcing the selection of clean samples from the entire feature space. Despite its simplicity, the proposed method can distinguish clean data from the corrupt ones more precisely and achieve state-of-the-art performance on image classification benchmarks, especially when the number of training samples is small and the noise rate is large. | Withdrawn |
ICLR.cc/2023/Conference | Benchmarking Offline Reinforcement Learning on Real-Robot Hardware | Learning policies from previously recorded data is a promising direction for real-world robotics tasks, as online learning is often infeasible. Dexterous manipulation in particular remains an open problem in its general form. The combination of offline reinforcement learning with large diverse datasets, however, has the potential to lead to a breakthrough in this challenging domain analogously to the rapid progress made in supervised learning in recent years. To coordinate the efforts of the research community toward tackling this problem, we propose a benchmark including: i) a large collection of data for offline learning from a dexterous manipulation platform on two tasks, obtained with capable RL agents trained in simulation; ii) the option to execute learned policies on a real-world robotic system and a simulation for efficient debugging. We evaluate prominent open-sourced offline reinforcement learning algorithms on the datasets and provide a reproducible experimental setup for offline reinforcement learning on real systems. | Accept: notable-top-25% |
ICLR.cc/2021/Conference | PURE: An Uncertainty-aware Recommendation Framework for Maximizing Expected Posterior Utility of Platform | Commercial recommendation can be regarded as an interactive process between the recommendation platform and its target users. One crucial problem for the platform is how to make full use of its advantages so as to maximize its utility, i.e., the commercial benefits from recommendation. In this paper, we propose a novel recommendation framework which effectively utilizes the information of user uncertainty over different item dimensions and explicitly takes into consideration the impact of display policy on user in order to achieve maximal expected posterior utility for the platform. We formulate the problem of deriving optimal policy to achieve maximal expected posterior utility as a constrained non-convex optimization problem and further propose an ADMM-based solution to derive an approximately optimal policy. Extensive experiments are conducted over data collected from a real-world recommendation platform and demonstrate the effectiveness of the proposed framework. Besides, we also adopt the proposed framework to conduct experiments with an intent to reveal how the platform achieves its commercial benefits. The results suggest that the platform should cater to the user's preference for item dimensions that the user prefers, while for item dimensions where the user is with high uncertainty, the platform can achieve more commercial benefits by recommending items with high utilities. | Reject |
ICLR.cc/2023/Conference | Explainable Machine Learning Predictions for the Long-term Performance of Brain-Computer Interfaces | Brain computer interfaces (BCIs) can decode neural signals to control assistive technologies such as robotic limbs for people with paralysis. Neural recordings from intracortical microelectrodes offer the spatiotemporal resolution (e.g., sortable units) necessary for complex tasks, such as controlling a robotic arm with multiple degrees of freedom. However, the quality of these signals decays over time despite many attempts to prolong their longevity. This decrease in long-term performance limits the implementation of this potentially beneficial technology. Predicting whether a channel will have sortable units across time would mitigate this issue and increase the utility of these devices by reducing uncertainty, yet to-date, no such methods exist. Similarly, it would be useful to understand how variables like time post-implantation, electrochemical characteristics, and electrode design impact the long-term quality of these signals. Here, we obtained longitudinal neural recordings and electrochemical data from freely behaving rats implanted with a custom designed microelectrode array with varying site areas, shank positions, and site depths. This dataset was used to develop an explainable artificial intelligence pipeline that predicts with high accuracy the presence of sortable units on a given channel and elucidates the most important factors leading to these predictions. Our pipeline was able to predict whether a channel will be active with an AUC of 0.79 (95% C.I. 0.73–0.86) on unseen data. The most important features of the model were experimental subject, time post-implantation, and a channel’s previous spike metrics. Electrode site depth was the most important electrode design variable. Our results demonstrate the feasibility of implementing explainable artificial intelligence pipelines for longitudinal BCI studies and support previous reports on how factors like time, inter-animal variability, and cortical depth impact long-term performance of BCIs. These results are an important step forward in improving efficient decoding performance and guiding device development, which stand to advance the field and benefit the lives of human BCI patients. | Reject |
ICLR.cc/2023/Conference | D4FT: A Deep Learning Approach to Kohn-Sham Density Functional Theory | Kohn-Sham Density Functional Theory (KS-DFT) has been traditionally solved by the Self-Consistent Field (SCF) method. Behind the SCF loop is the physics intuition of solving a system of non-interactive single-electron wave functions under an effective potential. In this work, we propose a deep learning approach to KS-DFT. First, in contrast to the conventional SCF loop, we propose to directly minimize the total energy by reparameterizing the orthogonal constraint as a feed-forward computation. We prove that such an approach has the same expressivity as the SCF method, yet reduces the computational complexity from O(N^4) to O(N^3). Second, the numerical integration which involves a summation over the quadrature grids can be amortized to the optimization steps. At each step, stochastic gradient descent (SGD) is performed with a sampled minibatch of the grids. Extensive experiments are carried out to demonstrate the advantage of our approach in terms of efficiency and stability. In addition, we show that our approach enables us to explore more complex neural-based wave functions. | Accept: notable-top-25% |
ICLR.cc/2021/Conference | Understanding and Mitigating Accuracy Disparity in Regression | With the widespread deployment of large-scale prediction systems in high-stakes domains, e.g., face recognition, criminal justice, etc., disparity on prediction accuracy between different demographic subgroups has called for fundamental understanding on the source of such disparity and algorithmic intervention to mitigate it. In this paper, we study the accuracy disparity problem in regression. To begin with, we first propose an error decomposition theorem, which decomposes the accuracy disparity into the distance between label populations and the distance between conditional representations, to help explain why such accuracy disparity appears in practice. Motivated by this error decomposition and the general idea of distribution alignment with statistical distances, we then propose an algorithm to reduce this disparity, and analyze its game-theoretic optima of the proposed objective function. We conduct experiments on four real-world datasets. The experimental results suggest that our proposed algorithms can effectively mitigate accuracy disparity while maintaining the predictive power of the regression models. | Reject |
ICLR.cc/2023/Conference | Algorithmic Determination of the Combinatorial Structure of the Linear Regions of ReLU Neural Networks | We algorithmically determine the regions and facets of all dimensions of the canonical polyhedral complex, the universal object into which a ReLU network decomposes its input space. We show that the locations of the vertices of the canonical polyhedral complex along with their signs with respect to layer maps determine the full facet structure across all dimensions. We present an algorithm which calculates this full combinatorial structure, making use of our theorems that the dual complex to the canonical polyhedral complex is cubical and it possesses a multiplication compatible with its facet structure. The resulting algorithm is numerically stable, polynomial time in the number of intermediate neurons, and obtains accurate information across all dimensions. This permits us to obtain, for example, the true topology of the decision boundaries of networks with low-dimensional inputs. We run empirics on such networks at initialization, finding that width alone does not increase observed topology, but width in the presence of depth does. | Reject |
ICLR.cc/2021/Conference | Contrastive Learning with Hard Negative Samples | We consider the question: how can you sample good negative examples for contrastive learning? We argue that, as with metric learning, learning contrastive representations benefits from hard negative samples (i.e., points that are difficult to distinguish from an anchor point). The key challenge toward using hard negatives is that contrastive methods must remain unsupervised, making it infeasible to adopt existing negative sampling strategies that use label information. In response, we develop a new class of unsupervised methods for selecting hard negative samples where the user can control the amount of hardness. A limiting case of this sampling results in a representation that tightly clusters each class, and pushes different classes as far apart as possible. The proposed method improves downstream performance across multiple modalities, requires only few additional lines of code to implement, and introduces no computational overhead.
| Accept (Poster) |
ICLR.cc/2019/Conference | UaiNets: From Unsupervised to Active Deep Anomaly Detection | This work presents a method for active anomaly detection which can be built upon existing deep learning solutions for unsupervised anomaly detection. We show that a prior needs to be assumed on what the anomalies are, in order to have performance guarantees in unsupervised anomaly detection. We argue that active anomaly detection has, in practice, the same cost of unsupervised anomaly detection but with the possibility of much better results. To solve this problem, we present a new layer that can be attached to any deep learning model designed for unsupervised anomaly detection to transform it into an active method, presenting results on both synthetic and real anomaly detection datasets. | Reject |
ICLR.cc/2023/Conference | Adversarial Collaborative Learning on Non-IID Features | Federated Learning (FL) has been a popular approach to enable collaborative learning on multiple parties without exchanging raw data. However, the model performance of FL may degrade a lot due to non-IID data. While many FL algorithms focus on non-IID labels, FL on non-IID features has largely been overlooked. Different from typical FL approaches, the paper proposes a new learning concept called ADCOL (Adversarial Collaborative Learning) for non-IID features. Instead of adopting the widely used model-averaging scheme, ADCOL conducts training in an adversarial way: the server aims to train a discriminator to distinguish the representations of the parties, while the parties aim to generate a common representation distribution. Our experiments on three tasks show that ADCOL achieves better performance than state-of-the-art FL algorithms on non-IID features. | Reject |
ICLR.cc/2021/Conference | End-to-End Egospheric Spatial Memory | Spatial memory, or the ability to remember and recall specific locations and objects, is central to autonomous agents' ability to carry out tasks in real environments. However, most existing artificial memory modules are not very adept at storing spatial information. We propose a parameter-free module, Egospheric Spatial Memory (ESM), which encodes the memory in an ego-sphere around the agent, enabling expressive 3D representations. ESM can be trained end-to-end via either imitation or reinforcement learning, and improves both training efficiency and final performance against other memory baselines on both drone and manipulator visuomotor control tasks. The explicit egocentric geometry also enables us to seamlessly combine the learned controller with other non-learned modalities, such as local obstacle avoidance. We further show applications to semantic segmentation on the ScanNet dataset, where ESM naturally combines image-level and map-level inference modalities. Through our broad set of experiments, we show that ESM provides a general computation graph for embodied spatial reasoning, and the module forms a bridge between real-time mapping systems and differentiable memory architectures. Implementation at: https://github.com/ivy-dl/memory. | Accept (Poster) |
ICLR.cc/2021/Conference | Time-varying Graph Representation Learning via Higher-Order Skip-Gram with Negative Sampling | Representation learning models for graphs are a successful family of techniques that project nodes into feature spaces that can be exploited by other machine learning algorithms. Since many real-world networks are inherently dynamic, with interactions among nodes changing over time, these techniques can be defined both for static and for time-varying graphs. Here, we show how the skip-gram embedding approach can be used to perform implicit tensor factorization on different tensor representations of time-varying graphs. We show that higher-order skip-gram with negative sampling (HOSGNS) is able to disentangle the role of nodes and time, with a small fraction of the number of parameters needed by other approaches. We empirically evaluate our approach using time-resolved face-to-face proximity data, showing that the learned representations outperform state-of-the-art methods when used to solve downstream tasks such as network reconstruction. Good performance on predicting the outcome of dynamical processes such as disease spreading shows the potential of this new method to estimate contagion risk, providing early risk awareness based on contact tracing data. | Reject |
ICLR.cc/2022/Conference | MULTI-LEVEL APPROACH TO ACCURATE AND SCALABLE HYPERGRAPH EMBEDDING | Many problems such as node classification and link prediction in network data
can be solved using graph embeddings, and a number of algorithms are known for constructing
such embeddings. However, it is difficult to use graphs to capture
non-binary relations such as communities of nodes. These kinds of
complex relations are expressed more naturally as hypergraphs.
While hypergraphs are a generalization of graphs, state-of-the-art graph
embedding techniques are not adequate for solving prediction and
classification tasks on large hypergraphs accurately in reasonable time.
In this paper, we introduce NetVec, a novel multi-level framework for scalable unsupervised hypergraph embedding, which outperforms state-of-the-art hypergraph embedding systems by up to
15% in accuracy.
We also show that NetVec is capable of generating high quality embeddings for real-world hypergraphs with millions of nodes and hyperedges in only a couple of minutes while existing hypergraph systems either fail for such large hypergraphs or may take days to produce the embeddings. | Withdrawn |
ICLR.cc/2021/Conference | A Learning Theoretic Perspective on Local Explainability | In this paper, we explore connections between interpretable machine learning and learning theory through the lens of local approximation explanations. First, we tackle the traditional problem of performance generalization and bound the test-time predictive accuracy of a model using a notion of how locally explainable it is. Second, we explore the novel problem of explanation generalization which is an important concern for a growing class of finite sample-based local approximation explanations. Finally, we validate our theoretical results empirically and show that they reflect what can be seen in practice. | Accept (Poster) |
ICLR.cc/2021/Conference | Redesigning the Classification Layer by Randomizing the Class Representation Vectors | Neural image classification models typically consist of two components. The first is an image encoder, which is responsible for encoding a given raw image into a representative vector. The second is the classification component, which is often implemented by projecting the representative vector onto target class vectors. The target class vectors, along with the rest of the model parameters, are estimated so as to minimize the loss function.
In this paper, we analyze how simple design choices for the classification layer affect the learning dynamics. We show that the standard cross-entropy training implicitly captures visual similarities between different classes, which might deteriorate accuracy or even prevents some models from converging. We propose to draw the class vectors randomly and set them as fixed during training, thus invalidating the visual similarities encoded in these vectors. We analyze the effects of keeping the class vectors fixed and show that it can increase the inter-class separability, intra-class compactness, and the overall model accuracy, while maintaining the robustness to image corruptions and the generalization of the learned concepts. | Reject |
ICLR.cc/2023/Conference | Generating Diverse Cooperative Agents by Learning Incompatible Policies | Training a robust cooperative agent requires diverse partner agents. However, obtaining those agents is difficult. Previous works aim to learn diverse behaviors by changing the state-action distribution of agents. But, without information about the task's goal, the diversified agents are not guided to find other important, albeit sub-optimal, solutions: the agents might learn only variations of the same solution. In this work, we propose to learn diverse behaviors via policy compatibility. Conceptually, policy compatibility measures whether policies of interest can coordinate effectively. We theoretically show that incompatible policies are not similar. Thus, policy compatibility—which has been used exclusively as a measure of robustness—can be used as a proxy for learning diverse behaviors. Then, we incorporate the proposed objective into a population-based training scheme to allow concurrent training of multiple agents. Additionally, we use state-action information to induce local variations of each policy. Empirically, the proposed method consistently discovers more solutions than baseline methods across various multi-goal cooperative environments. Finally, in multi-recipe Overcooked, we show that our method produces populations of behaviorally diverse agents, which enables generalist agents trained with such a population to be more robust.
See our project page at https://bit.ly/marl-lipo
| Accept: notable-top-25% |