conference
stringclasses 6
values | title
stringlengths 8
176
| abstract
stringlengths 228
5k
| decision
stringclasses 9
values |
|---|---|---|---|
ICLR.cc/2022/Conference | Nested Policy Reinforcement Learning for Clinical Decision Support | Off-policy reinforcement learning (RL) has proven to be a powerful framework for guiding agents' actions in environments with stochastic rewards and unknown or noisy state dynamics. In many real-world settings, these agents must operate in multiple environments, each with slightly different dynamics. For example, we may be interested in developing policies to guide medical treatment for patients with and without a given disease, or policies to navigate curriculum design for students with and without a learning disability. Here, we introduce nested policy fitted Q-iteration (NFQI), an RL framework that finds optimal policies in environments that exhibit such a structure. Our approach develops a nested $Q$-value function that takes advantage of the shared structure between two groups of observations from two separate environments while allowing their policies to be distinct from one another. We find that NFQI yields policies that rely on relevant features and perform at least as well as a policy that does not consider group structure. We demonstrate NFQI's performance using an OpenAI Gym environment and a clinical decision making RL task. Our results suggest that NFQI can develop policies that are better suited to many real-world clinical environments. | Reject |
ICLR.cc/2019/Conference | SNAS: stochastic neural architecture search | We propose Stochastic Neural Architecture Search (SNAS), an economical end-to-end solution to Neural Architecture Search (NAS) that trains neural operation parameters and architecture distribution parameters in same round of back-propagation, while maintaining the completeness and differentiability of the NAS pipeline. In this work, NAS is reformulated as an optimization problem on parameters of a joint distribution for the search space in a cell. To leverage the gradient information in generic differentiable loss for architecture search, a novel search gradient is proposed. We prove that this search gradient optimizes the same objective as reinforcement-learning-based NAS, but assigns credits to structural decisions more efficiently. This credit assignment is further augmented with locally decomposable reward to enforce a resource-efficient constraint. In experiments on CIFAR-10, SNAS takes less epochs to find a cell architecture with state-of-the-art accuracy than non-differentiable evolution-based and reinforcement-learning-based NAS, which is also transferable to ImageNet. It is also shown that child networks of SNAS can maintain the validation accuracy in searching, with which attention-based NAS requires parameter retraining to compete, exhibiting potentials to stride towards efficient NAS on big datasets. | Accept (Poster) |
ICLR.cc/2022/Conference | Permutation invariant graph-to-sequence model for template-free retrosynthesis and reaction prediction | Synthesis planning and reaction outcome prediction are two fundamental problems in computer-aided organic chemistry for which a variety of data-driven approaches have emerged. Natural language approaches that model each problem as a SMILES-to-SMILES translation lead to a simple end-to-end formulation, reduce the need for data preprocessing, and enable the use of well-optimized machine translation model architectures. However, SMILES representations are not an efficient representation for capturing information about molecular structure, as evidenced by the success of SMILES augmentation to boost empirical performance. Here, we describe a novel Graph2SMILES model that combines the power of Transformer models for text generation with the permutation invariance of molecular graph encoders. As an end-to-end architecture, Graph2SMILES can be used as a drop-in replacement for the Transformer in any task involving molecule(s)-to-molecule(s) transformations. In our encoder, an attention-augmented directed message passing neural network (D-MPNN) captures local chemical environments, and the global attention encoder allows for long-range and intermolecular interactions, enhanced by graph-aware positional embedding. Graph2SMILES improves the top-1 accuracy of the Transformer baselines by $1.7\%$ and $1.9\%$ for reaction outcome prediction on USPTO_480k and USPTO_STEREO datasets respectively, and by $9.8\%$ for one-step retrosynthesis on the USPTO_50k dataset. | Reject |
ICLR.cc/2022/Conference | GrASP: Gradient-Based Affordance Selection for Planning | Planning with a learned model is arguably a key component of intelligence. There are several challenges in realizing such a component in large-scale reinforcement learning (RL) problems. One such challenge is dealing effectively with continuous action spaces when using tree-search planning (e.g., it is not feasible to consider every action even at just the root node of the tree). In this paper we present a method for \emph{selecting} affordances useful for planning---for learning which small number of actions/options from a continuous space of actions/options to consider in the tree-expansion process during planning. We consider affordances that are goal-and-state-conditional mappings to actions/options as well as unconditional affordances that simply select actions/options available in all states. Our selection method is gradient based: we compute gradients through the planning procedure to update the parameters of the function that represents affordances. Our empirical work shows that it is feasible to learn to select both primitive-action and option affordances, and that simultaneously learning to select affordances and planning with a learned value-equivalent model can outperform model-free RL. | Reject |
ICLR.cc/2021/Conference | Predicting Infectiousness for Proactive Contact Tracing | The COVID-19 pandemic has spread rapidly worldwide, overwhelming manual contact tracing in many countries and resulting in widespread lockdowns for emergency containment. Large-scale digital contact tracing (DCT) has emerged as a potential solution to resume economic and social activity while minimizing spread of the virus. Various DCT methods have been proposed, each making trade-offs be-tween privacy, mobility restrictions, and public health. The most common approach, binary contact tracing (BCT), models infection as a binary event, informed only by an individual’s test results, with corresponding binary recommendations that either all or none of the individual’s contacts quarantine. BCT ignores the inherent uncertainty in contacts and the infection process, which could be used to tailor messaging to high-risk individuals, and prompt proactive testing or earlier warnings. It also does not make use of observations such as symptoms or pre-existing medical conditions, which could be used to make more accurate infectiousness predictions. In this paper, we use a recently-proposed COVID-19 epidemiological simulator to develop and test methods that can be deployed to a smartphone to locally and proactively predict an individual’s infectiousness (risk of infecting others) based on their contact history and other information, while respecting strong privacy constraints. Predictions are used to provide personalized recommendations to the individual via an app, as well as to send anonymized messages to the individual’s contacts, who use this information to better predict their own infectiousness, an approach we call proactive contact tracing (PCT). Similarly to other works, we find that compared to no tracing, all DCT methods tested are able to reduce spread of the disease and thus save lives, even at low adoption rates, strongly supporting a role for DCT methods in managing the pandemic. Further, we find a deep-learning based PCT method which improves over BCT for equivalent average mobility, suggesting PCT could help in safe re-opening and second-wave prevention. | Accept (Spotlight) |
ICLR.cc/2022/Conference | Picking up the pieces: separately evaluating supernet training and architecture selection | Differentiable Neural Architecture Search (NAS) has emerged as a simple and efficient method for the automated design of neural networks. Recent research has demonstrated improvements on various aspects on the original algorithm (DARTS), but comparative evaluation of these advances remains costly and difficult. We frame supernet NAS as a two-stage search, decoupling the training of the supernet from the extraction of a final design from the supernet. We propose a set of metrics which utilize benchmark data sets to evaluate each stage of the search process independently. We demonstrate two metrics measuring separately the quality of the supernet's shared weights and the quality of the learned sampling distribution, as well as corresponding statistics approximating the reliance of the second stage search on these components of the supernet. These metrics facilitate both more robust evaluation of NAS algorithms and provide practical method for designing complete NAS algorithms from separate supernet training and architecture selection techniques. | Reject |
ICLR.cc/2023/Conference | Fine-grained Few-shot Recognition by Deep Object Parsing | We propose a new method for fine-grained few-shot recognition via deep object parsing. In our framework, an object is made up of $K$ distinct parts and for each part, we learn a dictionary of templates, which is shared across all instances and categories. An object is parsed by estimating the locations of these $K$ parts and a set of active templates that can reconstruct the part features. We recognize test instances by comparing its active templates and the relative geometry of its part locations against those of the presented few-shot instances. Our method is end-to-end trainable to learn part templates on-top of a convolutional backbone. To combat visual distortions such as orientation, pose and size, we learn templates at multiple scales, and at test-time parse and match instances across these scales. We show that our method is competitive with the state-of-the-art, and by virtue of parsing enjoys interpretability as well. | Reject |
ICLR.cc/2023/Conference | Denoising Diffusion Error Correction Codes | Error correction code (ECC) is an integral part of the physical communication layer, ensuring reliable data transfer over noisy channels.
Recently, neural decoders have demonstrated their advantage over classical decoding techniques.
However, recent state-of-the-art neural decoders suffer from high complexity and lack the important iterative scheme characteristic of many legacy decoders.
In this work, we propose to employ denoising diffusion models for the soft decoding of linear codes at arbitrary block lengths.
Our framework models the forward channel corruption as a series of diffusion steps that can be reversed iteratively.
Three contributions are made: (i) a diffusion process suitable for the decoding setting is introduced, (ii) the neural diffusion decoder is conditioned on the number of parity errors, which indicates the level of corruption at a given step, (iii) a line search procedure based on the code's syndrome obtains the optimal reverse diffusion step size.
The proposed approach demonstrates the power of diffusion models for ECC and is able to achieve state-of-the-art accuracy, outperforming the other neural decoders by sizable margins, even for a single reverse diffusion step. | Accept: notable-top-25% |
ICLR.cc/2022/Conference | Sequential Covariate Shift Detection Using Classifier Two-Sample Tests | A standard assumption in supervised learning is that the training data and test data are from the same distribution. However, this assumption often fails to hold in practice, which can cause the learned model to perform poorly. We consider the problem of detecting covariate shift, where the covariate distribution shifts but the conditional distribution of labels given covariates remains the same. This problem can naturally be solved using a two-sample test--- i.e., test whether the current test distribution of covariates equals the training distribution of covariates. Our algorithm builds on classifier tests, which train a discriminator to distinguish train and test covariates, and then use the accuracy of this discriminator as a test statistic. A key challenge is that classifier tests assume given a fixed set of test covariates. In practice, test covariates often arrive sequentially over time---e.g., a self-driving car observes a stream of images while driving. Furthermore, covariate shift can occur multiple times--- i.e., shift and then shift back later or gradually shift over time. To address these challenges, our algorithm trains the discriminator online. Furthermore, it evaluates test accuracy using each new covariate before taking a gradient step; this strategy avoids constructing a held-out test set, which can reduce sample efficiency. We prove that this optimization preserves the correctness---i.e., our algorithm achieves a desired bound on the false positive rate. In our experiments, we show that our algorithm efficiently detects covariate shifts on ImageNet. | Reject |
ICLR.cc/2023/Conference | Individual Privacy Accounting with Gaussian Differential Privacy | Individual privacy accounting enables bounding differential privacy (DP) loss individually for each participant involved in the analysis. This can be informative as often the individual privacy losses are considerably smaller than those indicated by the DP bounds that are based on considering worst-case bounds at each data access. In order to account for the individual losses in a principled manner, we need a privacy accountant for adaptive compositions of mechanisms, where the loss incurred at a given data access is allowed to be smaller than the worst-case loss. This kind of analysis has been carried out for the Rényi differential privacy by Feldman and Zrnic (2021), however not yet for the so-called optimal privacy accountants. We make first steps in this direction by providing a careful analysis using the Gaussian differential privacy which gives optimal bounds for the Gaussian mechanism, one of the most versatile DP mechanisms. This approach is based on determining a certain supermartingale for the hockey-stick divergence and on extending the Rényi divergence-based fully adaptive composition results by Feldman and Zrnic (2021). We also consider measuring the individual $(\varepsilon,\delta)$-privacy losses using the so-called privacy loss distributions. Using the Blackwell theorem, we can then use the results of Feldman and Zrnic (2021) to construct an approximative individual $(\varepsilon,\delta)$-accountant. We also show how to speed up the FFT-based individual DP accounting using the Plancherel theorem. | Accept: poster |
ICLR.cc/2020/Conference | Deep Symbolic Superoptimization Without Human Knowledge | Deep symbolic superoptimization refers to the task of applying deep learning methods to simplify symbolic expressions. Existing approaches either perform supervised training on human-constructed datasets that defines equivalent expression pairs, or apply reinforcement learning with human-defined equivalent trans-formation actions. In short, almost all existing methods rely on human knowledge to define equivalence, which suffers from large labeling cost and learning bias, because it is almost impossible to define and comprehensive equivalent set. We thus propose HISS, a reinforcement learning framework for symbolic super-optimization that keeps human outside the loop. HISS introduces a tree-LSTM encoder-decoder network with attention to ensure tractable learning. Our experiments show that HISS can discover more simplification rules than existing human-dependent methods, and can learn meaningful embeddings for symbolic expressions, which are indicative of equivalence. | Accept (Poster) |
ICLR.cc/2022/Conference | Natural Attribute-based Shift Detection | Despite the impressive performance of deep networks in vision, language, and healthcare, unpredictable behaviors on samples from the distribution different than the training distribution cause severe problems in deployment. For better reliability of neural-network-based classifiers, we define a new task, natural attribute-based shift (NAS) detection, to detect the samples shifted from the training distribution by some natural attribute such as age of subjects or brightness of images. Using the natural attributes present in existing datasets, we introduce benchmark datasets in vision, language, and medical for NAS detection. Further, we conduct an extensive evaluation of prior representative out-of-distribution (OOD) detection methods on NAS datasets and observe an inconsistency in their performance. To understand this, we provide an analysis on the relationship between the location of NAS samples in the feature space and the performance of distance- and confidence-based OOD detection methods. Based on the analysis, we split NAS samples into three categories and further suggest a simple modification to the training objective to obtain an improved OOD detection method that is capable of detecting samples from all NAS categories. | Reject |
ICLR.cc/2021/Conference | Data augmentation as stochastic optimization | We present a theoretical framework recasting data augmentation as stochastic optimization for a sequence of time-varying proxy losses. This provides a unified language for understanding techniques commonly thought of as data augmentation, including synthetic noise and label-preserving transformations, as well as more traditional ideas in stochastic optimization such as learning rate and batch size scheduling. We then specialize our framework to study arbitrary augmentations in the context of a simple model (overparameterized linear regression). We extend in this setting the classical Monro-Robbins theorem to include augmentation and obtain rates of convergence, giving conditions on the learning rate and augmentation schedule under which augmented gradient descent converges. Special cases give provably good schedules for augmentation with additive noise, minibatch SGD, and minibatch SGD with noise. | Reject |
ICLR.cc/2021/Conference | Deep Convolution for Irregularly Sampled Temporal Point Clouds | We consider the problem of modeling the dynamics of continuous spatial-temporal processes represented by irregular samples through both space and time. Such processes occur in sensor networks, citizen science, multi-robot systems, and many others.
We propose a new deep model that is able to directly learn and predict over this irregularly sampled data, without voxelization, by leveraging a recent convolutional architecture for static point clouds. The model also easily incorporates the notion of multiple entities in the process. In particular, the model can flexibly answer prediction queries about arbitrary space-time points for different entities regardless of the distribution of the training or test-time data. We present experiments on real-world weather station data and battles between large armies in StarCraft II. The results demonstrate the model's flexibility in answering a variety of query types and demonstrate improved performance and efficiency compared to state-of-the-art baselines. | Reject |
ICLR.cc/2022/Conference | Generate, Annotate, and Learn: Generative Models Advance Self-Training and Knowledge Distillation | Semi-Supervised Learning (SSL) has seen success in many application domains, but this success often relies on the availability of task-specific unlabeled data. Knowledge distillation (KD) has enabled compressing deep networks, achieving the best results when distilling knowledge on fresh task-specific unlabeled examples. However, task-specific unlabeled data can be challenging to find, especially for NLP problems. We present a simple framework called "generate, annotate, and learn (GAL)" that uses unconditional language models to synthesize in-domain unlabeled data, helping advance SSL and KD on NLP and tabular tasks. To obtain strong task-specific generative models, we either fine-tune a large language model (LLM) on inputs from specific tasks, or prompt a LLM with a few input examples to generate more unlabeled examples. Then, we use existing classifiers to annotate generated unlabeled examples with pseudo labels, which are used as additional training data or as additional prompts. GAL improves prompt-based few-shot learning on several NLP tasks. It also yields a new state-of-the-art for 6-layer transformers on the GLUE leaderboard. Finally, self-training with GAL offers large gains on four tabular tasks from the UCI repository. | Reject |
ICLR.cc/2023/Conference | A CMDP-within-online framework for Meta-Safe Reinforcement Learning | Meta-reinforcement learning has widely been used as a learning-to-learn framework to solve unseen tasks with limited experience. However, the aspect of constraint violations has not been adequately addressed in the existing works, making their application restricted in real-world settings. In this paper, we study the problem of meta-safe reinforcement learning (meta-SRL) through the CMDP-within-online framework. We obtain task-averaged regret guarantees for the reward maximization (optimality gap) and constraint violations using gradient-based meta-learning and show that the task-averaged optimality gap and constraint satisfaction improve with task-similarity in the static environment, or task-relatedness in the changing environment. Several technical challenges arise when making this framework practical while still having strong theoretical guarantees. To address these challenges, we propose a meta-algorithm that performs inexact online learning on the upper bounds of intra-task optimality gap and constraint violations estimated by off-policy stationary distribution corrections. Furthermore, we enable the learning rates to be adapted for every task and extend our approach to settings with the dynamically changing task environments. Finally, experiments are conducted to demonstrate the effectiveness of our approach. The proposed theoretical framework is the first to handle the nonconvexity and stochastic nature of within-task CMDPs, while exploiting inter-task dependency for multi-task safe learning.
| Accept: notable-top-25% |
ICLR.cc/2021/Conference | ON NEURAL NETWORK GENERALIZATION VIA PROMOTING WITHIN-LAYER ACTIVATION DIVERSITY | During the last decade, neural networks have been intensively used to tackle various problems and they have often led to state-of-the-art results. These networks are composed of multiple jointly optimized layers arranged in a hierarchical structure. At each layer, the aim is to learn to extract hidden patterns needed to solve the problem at hand and forward it to the next layers. In the standard form, a neural network is trained with gradient-based optimization, where the errors are back-propagated from the last layer back to the first one. Thus at each optimization step, neurons at a given layer receive feedback from neurons belonging to higher layers of the hierarchy. In this paper, we propose to complement this traditional 'between-layer' feedback with additional 'within-layer' feedback to encourage diversity of the activations within the same layer. To this end, we measure the pairwise similarity between the outputs of the neurons and use it to model the layer's overall diversity. By penalizing similarities and promoting diversity, we encourage each neuron to learn a distinctive representation and, thus, to enrich the data representation learned within the layer and to increase the total capacity of the model. We theoretically study how the within-layer activation diversity affects the generalization performance of a neural network in a supervised context and we prove that increasing the diversity of hidden activations reduces the estimation error. In addition to the theoretical guarantees, we present an empirical study confirming that the proposed approach enhances the performance of neural networks. | Reject |
ICLR.cc/2021/Conference | Democratizing Evaluation of Deep Model Interpretability through Consensus | Deep learning interpretability tools, such as (Bau et al., 2017; Ribeiro et al., 2016; Smilkov et al., 2017), have been proposed to explain and visualize the ways that deep neural networks make predictions. The success of these methods highly relies on human subjective interpretations, i.e., the ground truth of interpretations, such as feature importance ranking or locations of visual objects, when evaluating the interpretability of the deep models on a specific task. For tasks that the ground truth of interpretations is not available, we propose a novel framework Consensus incorporating an ensemble of deep models as the committee for interpretability evaluation. Given any task/dataset, Consensus first obtains the interpretation results using existing tools, e.g., LIME (Ribeiro et al., 2016), for every model in the committee, then aggregates the results from the entire committee and approximates the “ground truth” of interpretations through voting. With such approximated ground truth, Consensus evaluates the interpretability of a model through matching its interpretation result and the approximated one, and ranks the matching scores together with committee members, so as to pursue the absolute and relative interpretability evaluation results. We carry out extensive experiments to validate Consensus on various datasets. The results show that Consensus can precisely identify the interpretability for a wide range of models on ubiquitous datasets that the ground truth is not available. Robustness analyses further demonstrate the advantage of the proposed framework to reach the consensus of interpretations through simple voting and evaluate the interpretability of deep models. Through the proposed Consensus framework, the interpretability evaluation has been democratized without the need of ground truth as criterion. | Reject |
ICLR.cc/2021/Conference | Search Data Structure Learning | In our modern world, an enormous amount of data surrounds us, and we are rarely interested in more than a handful of data points at once. It is like searching for needles in a haystack, and in many cases, there is no better algorithm than a random search, which might not be viable. Previously proposed algorithms for efficient database access are made for particular applications such as finding the min/max, finding all points within a range or finding the k-nearest neighbours. Consequently, there is a lack of versatility concerning what we can search when it comes to a gigantic database. In this work, we propose Search Data Structure Learning (SDSL), a generalization of the standard Search Data Structure (SDS) in which the machine has to learn how to search in the database. To evaluate approaches in this field, we propose a novel metric called Sequential Search Work Ratio (SSWR), a natural way of measuring a search's efficiency and quality. Finally, we inaugurate the field with the Efficient Learnable Binary Access (ELBA), a family of models for Search Data Structure Learning. It requires a means to train two parametric functions and a search data structure for binary codes. For the training, we developed a novel loss function, the F-beta Loss. For the SDS, we describe the Multi-Bernoulli Search (MBS), a novel approach for probabilistic binary codes. Finally, we exhibit the F-beta Loss and the MBS synergy by experimentally showing that it is at least twice as better than using the alternative loss functions of MIHash and HashNet and twenty times better than with another SDS based on the Hamming radius. | Reject |
ICLR.cc/2023/Conference | Thalamus: a brain-inspired algorithm for biologically-plausible continual learning and disentangled representations | Animals thrive in a constantly changing environment and leverage the temporal structure to learn well-factorized causal representations. In contrast, traditional neural networks suffer from forgetting in changing environments and many methods have been proposed to limit forgetting with different trade-offs. Inspired by the brain thalamocortical circuit, we introduce a simple algorithm that uses optimization at inference time to generate internal representations of the current task dynamically. The algorithm alternates between updating the model weights and a latent task embedding, allowing the agent to parse the stream of temporal experience into discrete events and organize learning about them. On a continual learning benchmark, it achieves competitive end average accuracy by mitigating forgetting, but importantly, the interaction between the weights dynamics and the latent dynamics organizes knowledge into flexible structures with a cognitive interface to control them. Tasks later in the sequence can be solved through knowledge transfer as they become reachable within the well-factorized latent space. The algorithm meets many of the desiderata of an ideal continually learning agent in open-ended environments, and its simplicity suggests fundamental computations in circuits with abundant feedback control loops such as the thalamocortical circuits in the brain | Accept: poster |
ICLR.cc/2023/Conference | PD-MORL: Preference-Driven Multi-Objective Reinforcement Learning Algorithm | Multi-objective reinforcement learning (MORL) approaches have emerged to tackle many real-world problems with multiple conflicting objectives by maximizing a joint objective function weighted by a preference vector. These approaches find fixed customized policies corresponding to preference vectors specified during training. However, the design constraints and objectives typically change dynamically in real-life scenarios. Furthermore, storing a policy for each potential preference is not scalable. Hence, obtaining a set of Pareto front solutions for the entire preference space in a given domain with a single training is critical. To this end, we propose a novel MORL algorithm that trains a single universal network to cover the entire preference space scalable to continuous robotic tasks. The proposed approach, Preference-Driven MORL (PD-MORL), utilizes the preferences as guidance to update the network parameters. It also employs a novel parallelization approach to increase sample efficiency. We show that PD-MORL achieves up to 25% larger hypervolume for challenging continuous control tasks and uses an order of magnitude fewer trainable parameters compared to prior approaches. | Accept: poster |
ICLR.cc/2020/Conference | Self-Supervised Policy Adaptation | We consider the problem of adapting an existing policy when the environment representation changes. Upon a change of the encoding of the observations the agent can no longer make use of its policy as it cannot correctly interpret the new observations. This paper proposes Greedy State Representation Learning (GSRL) to transfer the original policy by translating the environment representation back into its original encoding. To achieve this GSRL samples observations from both the environment and a dynamics model trained from prior experience. This generates pairs of state encodings, i.e., a new representation from the environment and a (biased) old representation from the forward model, that allow us to bootstrap a neural network model for state translation. Although early translations are unsatisfactory (as expected), the agent eventually learns a valid translation as it minimizes the error between expected and observed environment dynamics. Our experiments show the efficiency of our approach and that it translates the policy in considerably less steps than it would take to retrain the policy. | Reject |
ICLR.cc/2020/Conference | XLDA: Cross-Lingual Data Augmentation for Natural Language Inference and Question Answering | While natural language processing systems often focus on a single language, multilingual transfer learning has the potential to improve performance, especially for low-resource languages.
We introduce XLDA, cross-lingual data augmentation, a method that replaces a segment of the input text with its translation in another language. XLDA enhances performance of all 14 tested languages of the cross-lingual natural language inference (XNLI) benchmark. With improvements of up to 4.8, training with XLDA achieves state-of-the-art performance for Greek, Turkish, and Urdu. XLDA is in contrast to, and performs markedly better than, a more naive approach that aggregates examples in various languages in a way that each example is solely in one language. On the SQuAD question answering task, we see that XLDA provides a 1.0 performance increase on the English evaluation set. Comprehensive experiments suggest that most languages are effective as cross-lingual augmentors, that XLDA is robust to a wide range of translation quality, and that XLDA is even more effective for randomly initialized models than for pretrained models. | Reject |
ICLR.cc/2023/Conference | Learning Instance-Solution Operator For Optimal Control | Optimal control problems (OCPs) aim at finding a control function for a dynamical system such that a cost functional is optimized. These problems are central to physical system research in both academia and industry. In this paper, we propose a novel instance-solution operator learning perspective, which solves OCPs in a one-shot manner with no dependence on the explicit expression of dynamics or iterative optimization processes. The design is in principle endowed with substantial speedup in running time, and the model reusability is guaranteed by high-quality in- and out-of-distribution generalization. We theoretically validate the perspective by presenting the approximation bounds for the instance-solution operator learning. Extensive experiments on 6 physical systems verify the effectiveness and efficiency of our approach. The source code will be made publicly available. | Reject |
ICLR.cc/2019/Conference | Domain Generalization via Invariant Representation under Domain-Class Dependency | Learning domain-invariant representation is a dominant approach for domain generalization, where we need to build a classifier that is robust toward domain shifts induced by change of users, acoustic or lighting conditions, etc. However, prior domain-invariance-based methods overlooked the underlying dependency of classes (target variable) on source domains during optimization, which causes the trade-off between classification accuracy and domain-invariance, and often interferes with the domain generalization performance. This study first provides the notion of domain generalization under domain-class dependency and elaborates on the importance of considering the dependency by expanding the analysis of Xie et al. (2017). We then propose a method, invariant feature learning under optimal classifier constrains (IFLOC), which explicitly considers the dependency and maintains accuracy while improving domain-invariance. Specifically, the proposed method regularizes the representation so that it has as much domain information as the class labels, unlike prior methods that remove all domain information. Empirical validations show the superior performance of IFLOC to baseline methods, supporting the importance of the domain-class dependency in domain generalization and the efficacy of the proposed method for overcoming the issue. | Reject |
ICLR.cc/2022/Conference | Spherical Message Passing for 3D Molecular Graphs | We consider representation learning of 3D molecular graphs in which each atom is associated with a spatial position in 3D. This is an under-explored area of research, and a principled message passing framework is currently lacking. In this work, we conduct analyses in the spherical coordinate system (SCS) for the complete identification of 3D graph structures. Based on such observations, we propose the spherical message passing (SMP) as a novel and powerful scheme for 3D molecular learning. SMP dramatically reduces training complexity, enabling it to perform efficiently on large-scale molecules. In addition, SMP is capable of distinguishing almost all molecular structures, and the uncovered cases may not exist in practice. Based on meaningful physically-based representations of 3D information, we further propose the SphereNet for 3D molecular learning. Experimental results demonstrate that the use of meaningful 3D information in SphereNet leads to significant performance improvements in prediction tasks. Our results also demonstrate the advantages of SphereNet in terms of capability, efficiency, and scalability. | Accept (Poster) |
ICLR.cc/2022/Conference | Analytically Tractable Bayesian Deep Q-Learning | Reinforcement learning (RL) has gained increasing interest since the demonstration it was able to reach human performance on video game benchmarks using deep Q-learning (DQN). The current consensus of DQN for training neural networks (NNs) on such complex environments is to rely on gradient-descent optimization (GD). This consensus ignores the uncertainty of the NN's parameters which is a key aspect for the selection of an optimal action given a state. Although alternative Bayesian deep learning methods exist, most of them still rely on GD and numerical approximations, and they typically do not scale on complex benchmarks such as the Atari game environment. In this paper, we present how we can adapt the temporal difference Q-learning framework to make it compatible with the tractable approximate Gaussian inference (TAGI) which allows estimating the posterior distribution of NN's parameters using a closed-form analytical method. Throughout the experiments with on- and off-policy reinforcement learning approaches, we demonstrate that TAGI can reach a performance comparable to backpropagation-trained networks while using only half the number of hyperparameters, and without relying on GD or numerical approximations. | Reject |
ICLR.cc/2019/Conference | Exploiting Environmental Variation to Improve Policy Robustness in Reinforcement Learning | Conventional reinforcement learning rarely considers how the physical variations in the environment (eg. mass, drag, etc.) affect the policy learned by the agent. In this paper, we explore how changes in the environment affect policy generalization. We observe experimentally that, for each task we considered, there exists an optimal environment setting that results in the most robust policy that generalizes well to future environments. We propose a novel method to exploit this observation to develop robust actor policies, by automatically developing a sampling curriculum over environment settings to use in training. Ours is a model-free approach and experiments demonstrate that the performance of our method is on par with the best policies found by an exhaustive grid search, while bearing a significantly lower computational cost. | Reject |
ICLR.cc/2022/Conference | Online Ad Hoc Teamwork under Partial Observability | Autonomous agents often need to work together as a team to accomplish complex cooperative tasks. Due to privacy and other realistic constraints, agents might need to collaborate with previously unknown teammates on the fly. This problem is known as ad hoc teamwork, which remains a core research challenge. Prior works usually rely heavily on strong assumptions like full observability, fixed and predefined teammates' types. This paper relaxes these assumptions with a novel reinforcement learning framework called ODITS, which allows the autonomous agent to adapt to arbitrary teammates in an online fashion. Instead of limiting teammates into a finite set of predefined types, ODITS automatically learns latent variables of teammates' behaviors to infer how to cooperate with new teammates effectively. To overcome partial observability, we introduce an information-based regularizer to derive proxy representations of the learned variables from local observations. Extensive experimental results show that ODITS significantly outperforms various baselines in widely used ad hoc teamwork tasks. | Accept (Poster) |
ICLR.cc/2022/Conference | One for Many: an Instagram inspired black-box adversarial attack | It is well known that deep learning models are susceptible to adversarial attacks. To produce more robust and effective attacks, we propose a nested evolutionary algorithm able to produce multi-network (decision-based) black-box adversarial attacks based on Instagram inspired image filters. Due to the multi-network training, the system reaches a high transferability rate of attacks and, due to the composition of image filters, it is able to bypass standard detection mechanisms. Moreover, this kind of attack is semantically robust: our filter composition cannot be distinguished from any other filter composition used extensively every day to enhance images; this raises new security issues and challenges for real-world systems. Experimental results demonstrate that the method is also effective against
ensemble-adversarially trained models and it has a low cost in terms of queries to the victim model. | Reject |
ICLR.cc/2023/Conference | Motif-induced Graph Normalization | Graph Neural Networks (GNNs) have emerged as a powerful category of learning architecture for handling graph-structured data in the non-Euclidean domain. Despite their success, existing GNNs typically suffer from the insufficient expressive power bottlenecked by Weisfeiler-Lehman (WL) test, and meanwhile are prone to the over-smoothing situation with increasing layer numbers. In this paper, we strive to strengthen the discriminative capabilities of GNNs by devising a dedicated plug-and-play normalization scheme, termed as Motif-induced Normalization (MotifNorm), that explicitly considers the intra-connection information within each node-induced subgraph. To this end, we embed the motif-induced structural weights at the beginning and the end of the standard BatchNorm, as well as incorporate the graph instance-specific statistics for improved distinguishable capabilities. In the meantime, we provide the theoretical analysis to support that, with the proposed elaborated MotifNorm, an arbitrary GNNs is capable of more expressive abilities than the 1-WL test in distinguishing k-regular graphs. Furthermore, the proposed MotifNorm scheme is also exemplified to be able to alleviate the over-smoothing phenomenon. Experimental results on ten popular benchmarks across all the tasks of the graph-, node-, as well as link-level property predictions, demonstrate the effectiveness of the proposed method. Our code is made available in the supplementary material. | Reject |
ICLR.cc/2019/Conference | Robust Determinantal Generative Classifier for Noisy Labels and Adversarial Attacks | Large-scale datasets may contain significant proportions of noisy (incorrect) class labels, and it is well-known that modern deep neural networks poorly generalize from such noisy training datasets. In this paper, we propose a novel inference method, Deep Determinantal Generative Classifier (DDGC), which can obtain a more robust decision boundary under any softmax neural classifier pre-trained on noisy datasets. Our main idea is inducing a generative classifier on top of hidden feature spaces of the discriminative deep model. By estimating the parameters of generative classifier using the minimum covariance determinant estimator, we significantly improve the classification accuracy, with neither re-training of the deep model nor changing its architectures. In particular, we show that DDGC not only generalizes well from noisy labels, but also is robust against adversarial perturbations due to its large margin property. Finally, we propose the ensemble version ofDDGC to improve its performance, by investigating the layer-wise characteristics of generative classifier. Our extensive experimental results demonstrate the superiority of DDGC given different learning models optimized by various training techniques to handle noisy labels or adversarial samples. For instance, on CIFAR-10 dataset containing 45% noisy training labels, we improve the test accuracy of a deep model optimized by the state-of-the-art noise-handling training method from33.34% to 43.02%. | Reject |
ICLR.cc/2019/Conference | Hindsight policy gradients | A reinforcement learning agent that needs to pursue different goals across episodes requires a goal-conditional policy. In addition to their potential to generalize desirable behavior to unseen goals, such policies may also enable higher-level planning based on subgoals. In sparse-reward environments, the capacity to exploit information about the degree to which an arbitrary goal has been achieved while another goal was intended appears crucial to enable sample efficient learning. However, reinforcement learning agents have only recently been endowed with such capacity for hindsight. In this paper, we demonstrate how hindsight can be introduced to policy gradient methods, generalizing this idea to a broad class of successful algorithms. Our experiments on a diverse selection of sparse-reward environments show that hindsight leads to a remarkable increase in sample efficiency. | Accept (Poster) |
ICLR.cc/2022/Conference | Understanding Generalized Label Smoothing when Learning with Noisy Labels | Label smoothing (LS) is an arising learning paradigm that uses the positively weighted average of both the hard training labels and uniformly distributed soft labels. It was shown that LS serves as a regularizer for training data with hard labels and therefore improves the generalization of the model. Later it was reported LS even helps with improving robustness when learning with noisy labels. However, we observe that the advantage of LS vanishes when we operate in a high label noise regime. Puzzled by the observation, we proceeded to discover that several proposed learning-with-noisy-labels solutions in the literature instead relate more closely to $\textit{negative label smoothing}$ (NLS), which defines as using a negative weight to combine the hard and soft labels! We show that NLS differs substantially from LS in their achieved model confidence. To differentiate the two cases, we will call LS the positive label smoothing (PLS), and this paper unifies PLS and NLS into $\textit{generalized label smoothing}$ (GLS). We provide understandings for the properties of GLS when learning with noisy labels. Among other established properties, we theoretically show NLS is considered more beneficial when the label noise rates are high. We provide extensive experimental results on multiple benchmarks to support our findings too. | Reject |
ICLR.cc/2021/Conference | Diffeomorphic Template Transformers | In this paper we propose a spatial transformer network where the spatial transformations are limited to the group of diffeomorphisms. Diffeomorphic transformations are a kind of homeomorphism, which by definition preserve topology, a compelling property in certain applications.
We apply this diffemorphic spatial transformer to model the output of a neural network as a topology preserving mapping of a prior shape. By carefully choosing the prior shape we can enforce properties on the output of the network without requiring any changes to the loss function, such as smooth boundaries and a hard constraint on the number of connected components.
The diffeomorphic transformer networks outperform their non-diffeomorphic precursors when applied to learn data invariances in classification tasks. On a breast tissue segmentation task, we show that the approach is robust and flexible enough to deform simple artificial priors, such as Gaussian-shaped prior energies, into high-quality predictive probability densities. In addition to desirable topological properties, the segmentation maps have competitive quantitative fidelity compared to those obtained by direct estimation (i.e. plain U-Net). | Reject |
ICLR.cc/2022/Conference | Consistent Counterfactuals for Deep Models | Counterfactual examples are one of the most commonly-cited methods for explaining the predictions of machine learning models in key areas such as finance and medical diagnosis. Counterfactuals are often discussed under the assumption that the model on which they will be used is static, but in deployment models may be periodically retrained or fine-tuned. This paper studies the consistency of model prediction on counterfactual examples in deep networks under small changes to initial training conditions, such as weight initialization and leave-one-out variations in data, as often occurs during model deployment. We demonstrate experimentally that counterfactual examples for deep models are often inconsistent across such small changes, and that increasing the cost of the counterfactual, a stability-enhancing mitigation suggested by prior work in the context of simpler models, is not a reliable heuristic in deep networks. Rather, our analysis shows that a model's Lipschitz continuity around the counterfactual, along with confidence of its prediction, is key to its consistency across related models. To this end, we propose Stable Neighbor Search as a way to generate more consistent counterfactual explanations, and illustrate the effectiveness of this approach on several benchmark datasets. | Accept (Poster) |
ICLR.cc/2023/Conference | Modeling Temporal Data as Continuous Functions with Process Diffusion | Temporal data like time series are often observed at irregular intervals which is a challenging setting for the existing machine learning methods. To tackle this problem, we view such data as samples from some underlying continuous function. We then define a diffusion-based generative model that adds noise from a predefined stochastic process while preserving the continuity of the resulting underlying function. A neural network is trained to reverse this process which allows us to sample new realizations from the learned distribution. We define suitable stochastic processes as noise sources and introduce novel denoising and score-matching models on processes. Further, we show how to apply this approach to the multivariate probabilistic forecasting and imputation tasks. Through our extensive experiments, we demonstrate that our method outperforms previous models on synthetic and real-world datasets. | Reject |
ICLR.cc/2023/Conference | Koopman Neural Operator Forecaster for Time-series with Temporal Distributional Shifts | Temporal distributional shifts, with underlying dynamics changing over time, frequently occur in real-world time series and pose a fundamental challenge for deep neural networks (DNNs). In this paper, we propose a novel deep sequence model based on the Koopman theory for time series forecasting: Koopman Neural Forecaster (KNF) that leverages DNNs to learn the linear Koopman space and the coefficients of chosen measurement functions. KNF imposes appropriate inductive biases for improved robustness against distributional shifts, employing both a global operator to learn shared characteristics and a local operator to capture changing dynamics, as well as a specially-designed feedback loop to continuously update the learnt operators over time for rapidly varying behaviors. We demonstrate that KNF achieves superior performance compared to the alternatives, on multiple time series datasets that are shown to suffer from distribution shifts. | Accept: poster |
ICLR.cc/2019/Conference | Morpho-MNIST: Quantitative Assessment and Diagnostics for Representation Learning | Revealing latent structure in data is an active field of research, having introduced exciting technologies such as variational autoencoders and adversarial networks, and is essential to push machine learning towards unsupervised knowledge discovery. However, a major challenge is the lack of suitable benchmarks for an objective and quantitative evaluation of learned representations. To address this issue we introduce Morpho-MNIST, a framework that aims to answer: "to what extent has my model learned to represent specific factors of variation in the data?" We extend the popular MNIST dataset by adding a morphometric analysis enabling quantitative comparison of trained models, identification of the roles of latent variables, and characterisation of sample diversity. We further propose a set of quantifiable perturbations to assess the performance of unsupervised and supervised methods on challenging tasks such as outlier detection and domain adaptation. | Reject |
ICLR.cc/2023/Conference | Augmentation Component Analysis: Modeling Similarity via the Augmentation Overlaps | Self-supervised learning aims to learn a embedding space where semantically similar samples are close. Contrastive learning methods pull views of samples together and push different samples away, which utilizes semantic invariance of augmentation but ignores the relationship between samples. To better exploit the power of augmentation, we observe that semantically similar samples are more likely to have similar augmented views. Therefore, we can take the augmented views as a special description of a sample. In this paper, we model such a description as the augmentation distribution, and we call it augmentation feature. The similarity in augmentation feature reflects how much the views of two samples overlap and is related to their semantical similarity. Without computational burdens to explicitly estimate values of the augmentation feature, we propose Augmentation Component Analysis (ACA) with a contrastive-like loss to learn principal components and an on-the-fly projection loss to embed data. ACA equals an efficient dimension reduction by PCA and extracts low-dimensional embeddings, theoretically preserving the similarity of augmentation distribution between samples. Empirical results show that our method can achieve competitive results against various traditional contrastive learning methods on different benchmarks. | Accept: poster |
ICLR.cc/2021/Conference | Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels | We propose a simple data augmentation technique that can be applied to standard model-free reinforcement learning algorithms, enabling robust learning directly from pixels without the need for auxiliary losses or pre-training. The approach leverages input perturbations commonly used in computer vision tasks to transform input examples, as well as regularizing the value function and policy. Existing model-free approaches, such as Soft Actor-Critic (SAC), are not able to train deep networks effectively from image pixels. However, the addition of our augmentation method dramatically improves SAC’s performance, enabling it to reach state-of-the-art performance on the DeepMind control suite, surpassing model-based (Hafner et al., 2019; Lee et al., 2019; Hafner et al., 2018) methods and recently proposed contrastive learning (Srinivas et al., 2020). Our approach, which we dub DrQ: Data-regularized Q, can be combined with any model-free reinforcement learning algorithm. We further demonstrate this by applying it to DQN and significantly improve its data-efficiency on the Atari 100k benchmark. | Accept (Spotlight) |
ICLR.cc/2023/Conference | Where prior learning can and can't work in unsupervised inverse problems | Linear inverse problems consist in recovering a signal from its noisy observation in a lower dimensional space. Many popular resolution methods rely on data-driven algorithms that learn a prior from pairs of signals and observations to overcome the loss of information. However, these approaches are difficult, if not impossible, to adapt to unsupervised contexts -- where no ground truth data are available -- due to the need for learning from clean signals. This paper studies situations that do or do not allow learning a prior in unsupervised inverse problems. First, we focus on dictionary learning and point out that recovering the dictionary is unfeasible without constraints when the signal is observed through only one measurement operator. It can, however, be learned with multiple operators, given that they are diverse enough to span the whole signal space. Then, we study methods where weak priors are made available either through optimization constraints or deep learning architectures. We empirically emphasize that they perform better than hand-crafted priors only if they are adapted to the inverse problem. | Reject |
ICLR.cc/2023/Conference | Energy Transformer | Transformers have become the de facto models of choice in machine learning, typically leading to impressive performance on many applications. At the same time, the architectural development in the transformer world is mostly driven by empirical findings, and the theoretical understanding of their architectural building blocks is rather limited. In contrast, Dense Associative Memory models or Modern Hopfield Networks have a well-established theoretical foundation, but have not yet demonstrated truly impressive practical results. We propose a transformer architecture that replaces the sequence of feedforward transformer blocks with a single large Associative Memory model. Our novel architecture, called Energy Transformer (or ET for short), has many of the familiar architectural primitives that are often used in the current generation of transformers. However, it is not identical to the existing architectures. The sequence of transformer layers in ET is purposely designed to minimize a specifically engineered energy function, which is responsible for representing the relationships between the tokens. As a consequence of this computational principle, the attention in ET is different from the conventional attention mechanism. In this work, we introduce the theoretical foundations of ET, explore it's empirical capabilities using the image completion task, and obtain strong quantitative results on the graph anomaly detection task. | Reject |
ICLR.cc/2023/Conference | Active Image Indexing | Image copy detection and retrieval from large databases leverage two components. First, a neural network maps an image to a vector representation, that is relatively robust to various transformations of the image. Second, an efficient but approximate similarity search algorithm trades scalability (size and speed) against quality of the search, thereby introducing a source of error.
This paper improves the robustness of image copy detection with active indexing, that optimizes the interplay of these two components. We reduce the quantization loss of a given image representation by making imperceptible changes to the image before its release. The loss is back-propagated through the deep neural network back to the image, under perceptual constraints. These modifications make the image more retrievable.
Our experiments show that the retrieval and copy detection of activated images is significantly improved. For instance, activation improves by $+40\%$ the Recall1@1 on various image transformations, and for several popular indexing structures based on product quantization and locality sensitivity hashing. | Accept: poster |
ICLR.cc/2022/Conference | ZeroSARAH: Efficient Nonconvex Finite-Sum Optimization with Zero Full Gradient Computations | We propose ZeroSARAH -- a novel variant of the variance-reduced method SARAH (Nguyen et al., 2017) -- for minimizing the average of a large number of nonconvex functions $\frac{1}{n}\sum_{i=1}^{n}f_i(x)$. To the best of our knowledge, in this nonconvex finite-sum regime, all existing variance-reduced methods, including SARAH, SVRG, SAGA and their variants, need to compute the full gradient over all $n$ data samples at the initial point $x^0$, and then periodically compute the full gradient once every few iterations (for SVRG, SARAH and their variants). Note that SVRG, SAGA and their variants typically achieve weaker convergence results than variants of SARAH: $n^{2/3}/\epsilon^2$ vs. $n^{1/2}/\epsilon^2$. Thus we focus on the variant of SARAH. The proposed ZeroSARAH and its distributed variant D-ZeroSARAH are the \emph{first} variance-reduced algorithms which \emph{do not require any full gradient computations}, not even for the initial point. Moreover, for both standard and distributed settings, we show that ZeroSARAH and D-ZeroSARAH obtain new state-of-the-art convergence results, which can improve the previous best-known result (given by e.g., SPIDER, SARAH, and PAGE) in certain regimes. Avoiding any full gradient computations (which are time-consuming steps) is important in many applications as the number of data samples $n$ usually is very large. Especially in the distributed setting, periodic computation of full gradient over all data samples needs to periodically synchronize all clients/devices/machines, which may be impossible or unaffordable. Thus, we expect that ZeroSARAH/D-ZeroSARAH will have a practical impact in distributed and federated learning where full device participation is impractical. | Reject |
ICLR.cc/2022/Conference | SpSC: A Fast and Provable Algorithm for Sampling-Based GNN Training | Neighbor sampling is a commonly used technique for training Graph Neural Networks (GNNs) on large graphs. Previous work has shown that sampling-based GNN training can be considered as Stochastic Compositional Optimization (SCO) problems and can be better solved by SCO algorithms. However, we find that SCO algorithms are impractical for training GNNs on large graphs because they need to store the moving averages of the aggregated features of all nodes in the graph. The moving averages can easily exceed the GPU memory limit and even the CPU memory limit. In this work, we propose a variant of SCO algorithms with sparse moving averages for GNN training. By storing the moving averages in the most recent iterations, our algorithm only requires a fixed size buffer, regardless of the graph size. We show that our algorithm can achieve $O(\sqrt{1/K})$ convergence rate when the buffer size satisfies certain conditions. Our experiments validate our theoretical results and show that our algorithm outperforms the traditional Adam SGD for GNN training with a small memory overhead. | Reject |
ICLR.cc/2023/Conference | Scenario-based Question Answering with Interacting Contextual Properties | In the scenario-based Question Answering (QA) task, models are asked to find answers that are appropriate to the user scenarios associated with the question and identify information that is missing from the scenarios but is necessary for the answers to hold. Scenarios commonly include multiple properties of users, such as age, employment status, and income level for the question “How much can I claim from this benefit”. The properties relevant to a potential answer are given in a document, which will state conditions necessary for the answer to hold. Documents also may specify how conditions interact with each other, e.g. with text like “one of the conditions below must apply”. Although understanding the relationship between conditions is crucial for solving this challenging QA task, limited work has been done so far in modeling this. In this paper, we propose the T-Reasoner model, which solves this problem with three jointly learned modules: an entailment module which checks whether a condition has been satisfied by the scenario, a decoding module which locates eligible answers from documents, and a reasoning module which infers the relationship between conditions and performs a reasoning step to determine the logically consistent answers and identify missing conditions. T-Reasoner outperforms strong baselines on a synthetic scenario-based QA dataset and achieves a new state-of-the-art on two scenario-based QA benchmarks, outperforming the prior best models by 3-10 points. | Accept: poster |
ICLR.cc/2022/Conference | Rewardless Open-Ended Learning (ROEL) | Open-ended learning algorithms aim to automatically generate challenges and solutions to an unending sequence of learning opportunities. In Reinforcement Learning (RL) recent approaches to open-ended learning, such as Paired Open-Ended Trailblazer (POET), focus on collecting a diverse set of solutions based on the novelty of an agents pre-defined reward function. In many practical RL tasks defining an effective reward function a priori is often hard and can hinder an agents ability to explore many behaviors that could ultimately be more performant. In this work we combine open-ended learning with unsupervised reinforcement learning to train agents to learn a diverse set of complex skills. We propose a procedure to combine skill-discovery via mutual information, using the POET algorithm as an open-ended framework to teach agents increasingly complex groups of diverse skills. Experimentally we demonstrate this approach yields agents capable of demonstrating identifiable skills over a range of environments, that can be extracted and utilized to solve a variety of tasks. | Reject |
ICLR.cc/2022/Conference | Short optimization paths lead to good generalization | Optimization and generalization are two essential aspects of machine learning. In this paper, we propose a framework to connect optimization with generalization by analyzing the generalization error based on the length of optimization trajectory under the gradient flow algorithm after convergence. Through our approach, we show that, with a proper initialization, gradient flow converges following a short path with an explicit length estimate. Such an estimate induces a length-based generalization bound, showing that short optimization paths after convergence indicate good generalization. Our framework can be applied to broad settings. For example, we use it to obtain generalization estimates on three distinct machine learning models: underdetermined $\ell_p$ linear regression, kernel regression, and overparameterized two-layer ReLU neural networks. | Reject |
ICLR.cc/2023/Conference | Disentangled Feature Swapping Augmentation for Weakly Supervised Semantic Segmentation | Weakly supervised semantic segmentation utilizes a localization map obtained from a classifier to generate a pseudo-mask. However, classifiers utilize background cues to predict class labels because of a biased dataset consisting of images, in which specific objects frequently co-occur with certain backgrounds. Consequently, the classifier confuses the background with the target objects, resulting in inaccurate localization maps. To this end, we propose DisEntangled FeaTure swapping augmentation(DEFT) to prevent the classifier from being biased by a misleading correlation. Our method first disentangles the foreground and background features. Then, we randomly swap the disentangled features within mini-batches via a two-way process. These features contain various contexts that do not appear in the biased dataset, but the class relevant representation is preserved. In addition, we introduce training schemes to obtain further performance gains. Experimental results showed that when our augmentation was used in various weakly supervised semantic segmentation methods trained on the Pascal VOC 2012 dataset, the performance of the localization maps and pseudo-mask as well as the segmentation results improved. | Reject |
ICLR.cc/2023/Conference | AutoFHE: Automated Adaption of CNNs for Efficient Evaluation over FHE | Secure inference of deep convolutional neural networks (CNNs) was recently demonstrated under the RNS-CKKS fully homomorphic encryption (FHE) scheme. The state-of-the-art solution uses a high-order composite polynomial to approximate non-arithmetic ReLUs and refreshes zero-level ciphertext through bootstrapping. However, this solution suffers from prohibitively high latency, both due to the number of levels consumed by the polynomials ($47\%$) and the inference time consumed by bootstrapping operations ($70\%$). Furthermore, it requires a hand-crafted architecture for homomorphically evaluating CNNs by placing a bootstrapping operation after every Conv-BN layer. To accelerate CNNs on FHE and automatically design a homomorphic evaluation architecture, we propose AutoFHE: Automated adaption of CNNs for evaluation over FHE. AutoFHE exploits the varying sensitivity of approximate activations across different layers in a network and jointly evolves polynomial activations (EvoReLUs) and searches for placement of bootstrapping operations for evaluation under RNS-CKKS. The salient features of AutoFHE include: i) a multi-objective co-evolutionary (MOCoEv) search algorithm to maximize validation accuracy and minimize the number of bootstrapping operations, ii) a gradient-free search algorithm, R-CCDE, to optimize EvoReLU coefficients, and iii) polynomial-aware training (PAT) to fine-tune polynomial-only CNNs for one epoch to adapt trainable weights to EvoReLUs. We demonstrate the efficacy of AutoFHE through the evaluation of ResNets on CIFAR-10 and CIFAR-100 under RNS-CKKS. Experimental results on CIFAR-10 indicate that in comparison to the state-of-the-art solution, AutoFHE reduces inference time (50 images on 50 threads) by 1,000 seconds and amortized inference time (per image) by $28\%$ and $17\%$ for ResNet-20 and ResNet-32, respectively. | Reject |
ICLR.cc/2023/Conference | Comfort Zone: A Vicinal Distribution for Regression Problems | Domain-dependent data augmentation methods generate artificial samples using transformations suited for the underlying data domain, for example rotations on images and time warping on time series data. However, domain-independent approaches, e.g. mixup, are applicable to various data modalities, and as such they are general and versatile. While mixup-based techniques are used extensively in classification problems, their effect on regression tasks is somewhat less explored. To bridge this gap, we study the problem of domain-independent augmentation for regression, and we introduce comfort-zone: a new data-driven, domain-independent data augmentation method. Essentially, our approach samples new examples from the tangent planes of the train distribution. Augmenting data in this way aligns with the network tendency towards capturing the dominant features of its input signals. Evaluating comfort-zone on regression and time series forecasting benchmarks, we show that it improves the generalization of several neural architectures. We also find that mixup and noise injection are less effective in comparison to comfort-zone. | Reject |
ICLR.cc/2022/Conference | Graph-Guided Network for Irregularly Sampled Multivariate Time Series | In many domains, including healthcare, biology, and climate science, time series are irregularly sampled with varying time intervals between successive readouts and different subsets of variables (sensors) observed at different time points. Here, we introduce RAINDROP, a graph neural network that embeds irregularly sampled and multivariate time series while also learning the dynamics of sensors purely from observational data. RAINDROP represents every sample as a separate sensor graph and models time-varying dependencies between sensors with a novel message passing operator. It estimates the latent sensor graph structure and leverages the structure together with nearby observations to predict misaligned readouts. This model can be interpreted as a graph neural network that sends messages over graphs that are optimized for capturing time-varying dependencies among sensors. We use RAINDROP to classify time series and interpret temporal dynamics on three healthcare and human activity datasets. RAINDROP outperforms state-of-the-art methods by up to 11.4% (absolute F1-score points), including techniques that deal with irregular sampling using fixed discretization and set functions. RAINDROP shows superiority in diverse setups, including challenging leave-sensor-out settings. | Accept (Poster) |
ICLR.cc/2021/Conference | Adversarial Boot Camp: label free certified robustness in one epoch | Machine learning models are vulnerable to adversarial attacks. One approach to addressing this vulnerability is certification, which focuses on models that are guaranteed to be robust for a given perturbation size. A drawback of recent certified models is that they are stochastic: they require multiple computationally expensive model evaluations with random noise added to a given image. In our work, we present a deterministic certification approach which results in a certifiably robust model. This approach is based on an equivalence between training with a particular regularized loss, and the expected values of Gaussian averages. We achieve certified models on ImageNet-1k by retraining a model with this loss for one epoch without the use of label information. | Reject |
ICLR.cc/2022/Conference | Finding General Equilibria in Many-Agent Economic Simulations using Deep Reinforcement Learning | Real economies can be seen as a sequential imperfect-information game with many heterogeneous, interacting strategic agents of various agent types, such as consumers, firms, and governments. Dynamic general equilibrium models are common economic tools to model the economic activity, interactions, and outcomes in such systems. However, existing analytical and computational methods struggle to find explicit equilibria when all agents are strategic and interact, while joint learning is unstable and challenging. Amongst others, a key reason is that the actions of one economic agent may change the reward function of another agent, e.g., a consumer's expendable income changes when firms change prices or governments change taxes. We show that multi-agent deep reinforcement learning (RL) can discover stable solutions that are $\epsilon$-Nash equilibria for a meta-game over agent types, in economic simulations with many agents, through the use of structured learning curricula and efficient GPU-only simulation and training.Conceptually, our approach is more flexible and does not need unrealistic assumptions, e.g., market clearing, that are commonly used for analytical tractability. Our GPU implementation enables training and analyzing economies with a large number of agents within reasonable time frames, e.g., training completes within a day. We demonstrate our approach in real-business-cycle models, a representative family of DGE models, with 100 worker-consumers, 10 firms, and a government who taxes and redistributes. We validate the learned meta-game $\epsilon$-Nash equilibria through approximate best-response analyses, show that RL policies align with economic intuitions, and that our approach is constructive, e.g., by explicitly learning a spectrum of meta-game $\epsilon$-Nash equilibria in open economic models. | Reject |
ICLR.cc/2023/Conference | MultiViz: Towards Visualizing and Understanding Multimodal Models | The promise of multimodal models for real-world applications has inspired research in visualizing and understanding their internal mechanics with the end goal of empowering stakeholders to visualize model behavior, perform model debugging, and promote trust in machine learning models. However, modern multimodal models are typically black-box neural networks, which makes it challenging to understand their internal mechanics. How can we visualize the internal modeling of multimodal interactions in these models? Our paper aims to fill this gap by proposing MultiViz, a method for analyzing the behavior of multimodal models by scaffolding the problem of interpretability into 4 stages: (1) unimodal importance: how each modality contributes towards downstream modeling and prediction, (2) cross-modal interactions: how different modalities relate with each other, (3) multimodal representations: how unimodal and cross-modal interactions are represented in decision-level features, and (4) multimodal prediction: how decision-level features are composed to make a prediction. MultiViz is designed to operate on diverse modalities, models, tasks, and research areas. Through experiments on 8 trained models across 6 real-world tasks, we show that the complementary stages in MultiViz together enable users to (1) simulate model predictions, (2) assign interpretable concepts to features, (3) perform error analysis on model misclassifications, and (4) use insights from error analysis to debug models. MultiViz is publicly available, will be regularly updated with new interpretation tools and metrics, and welcomes inputs from the community. | Accept: poster |
ICLR.cc/2021/Conference | Parallel Training of Deep Networks with Local Updates | Deep learning models trained on large data sets have been widely successful in both vision and language domains. As state-of-the-art deep learning architectures have continued to grow in parameter count so have the compute budgets and times required to train them, increasing the need for compute-efficient methods that parallelize training. Two common approaches to parallelize the training of deep networks have been data and model parallelism. While useful, data and model parallelism suffer from diminishing returns in terms of compute efficiency for large batch sizes. In this paper, we investigate how to continue scaling compute efficiently beyond the point of diminishing returns for large batches through local parallelism, a framework which parallelizes training of individual layers in deep networks by replacing global backpropagation with truncated layer-wise backpropagation. Local parallelism enables fully asynchronous layer-wise parallelism with a low memory footprint, and requires little communication overhead compared with model parallelism. We show results in both vision and language domains across a diverse set of architectures, and find that local parallelism is particularly effective in the high-compute regime. | Reject |
ICLR.cc/2020/Conference | Learning Neural Causal Models from Unknown Interventions | Meta-learning over a set of distributions can be interpreted as learning different types of parameters corresponding to short-term vs long-term aspects of the mechanisms underlying the generation of data. These are respectively captured by quickly-changing \textit{parameters} and slowly-changing \textit{meta-parameters}. We present a new framework for meta-learning causal models where the relationship between each variable and its parents is modeled by a neural network, modulated by structural meta-parameters which capture the overall topology of a directed graphical model. Our approach avoids a discrete search over models in favour of a continuous optimization procedure. We study a setting where interventional distributions are induced as a result of a random intervention on a single unknown variable of an unknown ground truth causal model, and the observations arising after such an intervention constitute one meta-example. To disentangle the slow-changing aspects of each conditional from the fast-changing adaptations to each intervention, we parametrize the neural network into fast parameters and slow meta-parameters. We introduce a meta-learning objective that favours solutions \textit{robust} to frequent but sparse interventional distribution change, and which generalize well to previously unseen interventions. Optimizing this objective is shown experimentally to recover the structure of the causal graph. Finally, we find that when the learner is unaware of the intervention variable, it is able to infer that information, improving results further and focusing the parameter and meta-parameter updates where needed. | Reject |
ICLR.cc/2023/Conference | More ConvNets in the 2020s: Scaling up Kernels Beyond 51x51 using Sparsity | Transformers have quickly shined in the computer vision world since the emergence of Vision Transformers (ViTs). The dominant role of convolutional neural networks (CNNs) seems to be challenged by increasingly effective transformer-based models. Very recently, a couple of advanced convolutional models strike back with large kernels motivated by the local-window attention mechanism, showing appealing performance and efficiency. While one of them, i.e. RepLKNet, impressively manages to scale the kernel size to 31x31 with improved performance, the performance starts to saturate as the kernel size continues growing, compared to the scaling trend of advanced ViTs such as Swin Transformer. In this paper, we explore the possibility of training extreme convolutions larger than 31x31 and test whether the performance gap can be eliminated by strategically enlarging convolutions. This study ends up with a recipe for applying extremely large kernels from the perspective of sparsity, which can smoothly scale up kernels to 61x61 with better performance. Built on this recipe, we propose Sparse Large Kernel Network (SLaK), a pure CNN architecture equipped with sparse factorized 51x51 kernels that can perform on par with or better than state-of-the-art hierarchical Transformers and modern ConvNet architectures like ConvNeXt and RepLKNet, on ImageNet classification as well as a wide range of downstream tasks including semantic segmentation on ADE20K, object detection on PASCAL VOC 2007, and object detection/segmentation on MS COCO. Codes are available at https://github.com/VITA-Group/SLaK. | Accept: poster |
ICLR.cc/2020/Conference | Diving into Optimization of Topology in Neural Networks | Seeking effective networks has become one of the most crucial and practical areas in deep learning. The architecture of a neural network can be represented as a directed acyclic graph, whose nodes denote transformation of layers and edges represent information flow. Despite the selection of \textit{micro} node operations, \textit{macro} connections among the whole network, noted as \textit{topology}, largely affects the optimization process. We first rethink the residual connections via a new \textit{topological view} and observe the benefits provided by dense connections to the optimization. Motivated by which, we propose an innovation method to optimize the topology of a neural network. The optimization space is defined as a complete graph, through assigning learnable weights which reflect the importance of connections, the optimization of topology is transformed into learning a set of continuous variables of edges. To extend the optimization to larger search spaces, a new series of networks,
named as TopoNet, are designed. To further focus on critical edges and promote generalization ablity in dense topologies, auxiliary sparsity constraint is adopted to constrain the distribution of edges. Experiments on classical networks prove the effectiveness of the optimization of topology. Experiments with TopoNets further verify both availability and transferability of the proposed method in
different tasks e.g. image classification, object detection and face recognition. | 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/2020/Conference | Self-Supervised GAN Compression | Deep learning's success has led to larger and larger models to handle more and more complex tasks; trained models can contain millions of parameters. These large models are compute- and memory-intensive, which makes it a challenge to deploy them with minimized latency, throughput, and storage requirements. Some model compression methods have been successfully applied on image classification and detection or language models, but there has been very little work compressing generative adversarial networks (GANs) performing complex tasks. In this paper, we show that a standard model compression technique, weight pruning, cannot be applied to GANs using existing methods. We then develop a self-supervised compression technique which uses the trained discriminator to supervise the training of a compressed generator. We show that this framework has a compelling performance to high degrees of sparsity, generalizes well to new tasks and models, and enables meaningful comparisons between different pruning granularities. | Reject |
ICLR.cc/2020/Conference | RTC-VAE: HARNESSING THE PECULIARITY OF TOTAL CORRELATION IN LEARNING DISENTANGLED REPRESENTATIONS | In the problem of unsupervised learning of disentangled representations, one of the promising methods is to penalize the total correlation of sampled latent vari-ables. Unfortunately, this well-motivated strategy often fail to achieve disentanglement due to a problematic difference between the sampled latent representation and its corresponding mean representation. We provide a theoretical explanation that low total correlation of sample distribution cannot guarantee low total correlation of the mean representation. We prove that for the mean representation of arbitrarily high total correlation, there exist distributions of latent variables of abounded total correlation. However, we still believe that total correlation could be a key to the disentanglement of unsupervised representative learning, and we propose a remedy, RTC-VAE, which rectifies the total correlation penalty. Experiments show that our model has a more reasonable distribution of the mean representation compared with baseline models, e.g.,β-TCVAE and FactorVAE. | Reject |
ICLR.cc/2021/Conference | SAFENet: A Secure, Accurate and Fast Neural Network Inference | The advances in neural networks have driven many companies to provide prediction services to users in a wide range of applications. However, current prediction systems raise privacy concerns regarding the user's private data. A cryptographic neural network inference service is an efficient way to allow two parties to execute neural network inference without revealing either party’s data or model. Nevertheless, existing cryptographic neural network inference services suffer from huge running latency; in particular, the latency of communication-expensive cryptographic activation function is 3 orders of magnitude higher than plaintext-domain activation function. And activations are the necessary components of the modern neural networks. Therefore, slow cryptographic activation has become the primary obstacle of efficient cryptographic inference.
In this paper, we propose a new technique, called SAFENet, to enable a Secure, Accurate and Fast nEural Network inference service. To speedup secure inference and guarantee inference accuracy, SAFENet includes channel-wise activation approximation with multiple-degree options. This is implemented by keeping the most useful activation channels and replacing the remaining, less useful, channels with various-degree polynomials. SAFENet also supports mixed-precision activation approximation by automatically assigning different replacement ratios to various layer; further increasing the approximation ratio and reducing inference latency. Our experimental results show SAFENet obtains the state-of-the-art inference latency and performance, reducing latency by $38\% \sim 61\%$ or improving accuracy by $1.8\% \sim 4\%$ over prior techniques on various encrypted datasets. | Accept (Poster) |
ICLR.cc/2023/Conference | Edgeformers: Graph-Empowered Transformers for Representation Learning on Textual-Edge Networks | Edges in many real-world social/information networks are associated with rich text information (e.g., user-user communications or user-product reviews). However, mainstream network representation learning models focus on propagating and aggregating node attributes, lacking specific designs to utilize text semantics on edges. While there exist edge-aware graph neural networks, they directly initialize edge attributes as a feature vector, which cannot fully capture the contextualized text semantics of edges. In this paper, we propose Edgeformers, a framework built upon graph-enhanced Transformers, to perform edge and node representation learning by modeling texts on edges in a contextualized way. Specifically, in edge representation learning, we inject network information into each Transformer layer when encoding edge texts; in node representation learning, we aggregate edge representations through an attention mechanism within each node’s ego-graph. On five public datasets from three different domains, Edgeformers consistently outperform state-of-the-art baselines in edge classification and link prediction, demonstrating the efficacy in learning edge and node representations, respectively. | Accept: poster |
ICLR.cc/2022/Conference | Adam is no better than normalized SGD: Dissecting how adaptivity improves GAN performance | Adaptive methods are widely used for training generative adversarial networks (GAN). While there has been some work to pinpoint the marginal value of adaptive methods in minimization problems, it remains unclear why it is still the method of choice for GAN training. This paper formally studies how adaptive methods help performance in GANs. First, we dissect Adam---the most popular adaptive method for GAN training---by comparing with SGDA the direction and the norm of its update vector. We empirically show that SGDA with the same vector norm as Adam reaches similar or even better performance than the latter. This empirical study encourages us to consider normalized stochastic gradient descent ascent (nSGDA) as a simpler alternative to Adam. We then propose a synthetic theoretical framework to understand why nSGDA yields better performance than SGDA for GANs. In that situation, we prove that a GAN trained with nSGDA provably recovers all the modes of the true distribution. In contrast, the same networks trained with SGDA (and any learning rate configuration) suffers from mode collapsing. The critical insight in our analysis is that normalizing the gradients forces the discriminator and generator to update at the same pace. We empirically show the competitive performance of nSGDA on real-world datasets.
| Reject |
ICLR.cc/2018/Conference | Graph2Seq: Scalable Learning Dynamics for Graphs | Neural networks are increasingly used as a general purpose approach to learning algorithms over graph structured data. However, techniques for representing graphs as real-valued vectors are still in their infancy. Recent works have proposed several approaches (e.g., graph convolutional networks), but as we show in this paper, these methods have difficulty generalizing to large graphs. In this paper we propose Graph2Seq, an embedding framework that represents graphs as an infinite time-series. By not limiting the representation to a fixed dimension, Graph2Seq naturally scales to graphs of arbitrary size. Moreover, through analysis of a formal computational model we show that an unbounded sequence is necessary for scalability. Graph2Seq is also reversible, allowing full recovery of the graph structure from the sequence. Experimental evaluations of Graph2Seq on a variety of combinatorial optimization problems show strong generalization and strict improvement over state of the art. | Reject |
ICLR.cc/2020/Conference | Address2vec: Generating vector embeddings for blockchain analytics | Bitcoin is a virtual coinage system that enables users to trade virtually free of a central trusted authority. All transactions on the Bitcoin blockchain are publicly available for viewing, yet as Bitcoin is built mainly for security it’s original structure does not allow for direct analysis of address transactions.
Existing analysis methods of the Bitcoin blockchain can be complicated, computationally expensive or inaccurate. We propose a computationally efficient model to analyze bitcoin blockchain addresses and allow for their use with existing machine learning algorithms. We compare our approach against Multi Level Sequence Learners (MLSLs), one of the best performing models on bitcoin address data. | Reject |
ICLR.cc/2022/Conference | Fair Node Representation Learning via Adaptive Data Augmentation | Node representation learning has demonstrated its efficacy for various applications on graphs, which leads to increasing attention towards the area. However, fairness is a largely under-explored territory within the field, which may lead to biased results towards underrepresented groups in ensuing tasks. To this end, this work theoretically explains the sources of bias in node representations obtained via Graph Neural Networks (GNNs). Our analysis reveals that both nodal features and graph structure lead to bias in the obtained representations. Building upon the analysis, fairness-aware data augmentation frameworks on nodal features and graph structure are developed to reduce the intrinsic bias. Our analysis and proposed schemes can be readily employed to enhance the fairness of various GNN-based learning mechanisms. Extensive experiments on node classification and link prediction are carried out over real networks in the context of graph contrastive learning. Comparison with multiple benchmarks demonstrates that the proposed augmentation strategies can improve fairness in terms of statistical parity and equal opportunity, while providing comparable utility to state-of-the-art contrastive methods. | Reject |
ICLR.cc/2020/Conference | Adversarial Robustness Against the Union of Multiple Perturbation Models | Owing to the susceptibility of deep learning systems to adversarial attacks, there has been a great deal of work in developing (both empirically and certifiably) robust classifiers, but the vast majority has defended against single types of attacks. Recent work has looked at defending against multiple attacks, specifically on the MNIST dataset, yet this approach used a relatively complex architecture, claiming that standard adversarial training can not apply because it "overfits" to a particular norm. In this work, we show that it is indeed possible to adversarially train a robust model against a union of norm-bounded attacks, by using a natural generalization of the standard PGD-based procedure for adversarial training to multiple threat models. With this approach, we are able to train standard architectures which are robust against l_inf, l_2, and l_1 attacks, outperforming past approaches on the MNIST dataset and providing the first CIFAR10 network trained to be simultaneously robust against (l_inf, l_2, l_1) threat models, which achieves adversarial accuracy rates of (47.6%, 64.3%, 53.4%) for (l_inf, l_2, l_1) perturbations with epsilon radius = (0.03,0.5,12). | Reject |
ICLR.cc/2023/Conference | Is Model Ensemble Necessary? Model-based RL via a Single Model with Lipschitz Regularized Value Function | Probabilistic dynamics model ensemble is widely used in existing model-based reinforcement learning methods as it outperforms a single dynamics model in both asymptotic performance and sample efficiency. In this paper, we provide both practical and theoretical insights on the empirical success of the probabilistic dynamics model ensemble through the lens of Lipschitz continuity. We find that, for a value function, the stronger the Lipschitz condition is, the smaller the gap between the true dynamics- and learned dynamics-induced Bellman operators is, thus enabling the converged value function to be closer to the optimal value function. Hence, we hypothesize that the key functionality of the probabilistic dynamics model ensemble is to regularize the Lipschitz condition of the value function using generated samples. To validate this hypothesis, we devise two practical robust training mechanisms through computing the adversarial noise and regularizing the value network’s spectral norm to directly regularize the Lipschitz condition of the value functions. Empirical results show that combined with our mechanisms, model-based RL algorithms with a single dynamics model outperform those with ensemble of the probabilistic dynamics models. These findings not only support the theoretical insight, but also provide a practical solution for developing computationally efficient model-based RL algorithms. | Accept: poster |
ICLR.cc/2023/Conference | Best Possible Q-Learning | Fully decentralized learning, where the global information, i.e., the actions of other agents, is inaccessible, is a fundamental challenge in multi-agent reinforcement learning. However, the convergence and optimality of most decentralized algorithms are not theoretically guaranteed, since the transition probabilities are non-stationary as all agents are updating policies simultaneously. To tackle this challenge, we propose \textit{best possible operator}, a novel decentralized operator, and prove that the policies of agents will converge to the optimal joint policy if each agent independently updates its individual state-action value by the operator. Further, to make the update more efficient and practical, we simplify the operator and prove that the convergence and optimality still hold with the simplified one. By instantiating the simplified operator, the derived fully decentralized algorithm, best possible Q-learning (BQL), does not suffer from non-stationarity. Empirically, we show that BQL achieves remarkable improvement over baselines in a variety of cooperative multi-agent tasks. | Reject |
ICLR.cc/2023/Conference | Is a Caption Worth a Thousand Images? A Study on Representation Learning | The development of CLIP [Radford et al., 2021] has sparked a debate on whether adding language supervision can yield vision models with more transferable representations than traditional image-only methods. Our work studies this question through a carefully controlled comparison of two approaches, in terms of their ability to learn representations that generalize to downstream classification tasks. We find that when the pre-training data meets certain criteria---it is sufficiently large and contains descriptive captions with low variability----image-only methods do not match CLIP's performance even when they are trained with more image data. However, contrary to what one might expect, there are practical settings in which these criteria are not met, wherein added supervision through captions is actually detrimental.
Motivated by our findings, we devise simple data and algorithmic interventions to improve the transfer performance of CLIP-style models. | Accept: poster |
ICLR.cc/2020/Conference | Deep Audio Prior | Deep convolutional neural networks are known to specialize in distilling compact and robust prior from a large amount of data. We are interested in applying deep networks in the absence of training dataset. In this paper, we introduce deep audio prior (DAP) which leverages the structure of a network and the temporal information in a single audio file. Specifically, we demonstrate that a randomly-initialized neural network can be used with carefully designed audio prior to tackle challenging audio problems such as universal blind source separation, interactive audio editing, audio texture synthesis, and audio co-separation.
To understand the robustness of the deep audio prior, we construct a benchmark dataset Universal-150 for universal sound source separation with a diverse set of sources. We show superior audio results than previous work on both qualitatively and quantitative evaluations. We also perform thorough ablation study to validate our design choices. | Reject |
ICLR.cc/2021/Conference | Randomized Ensembled Double Q-Learning: Learning Fast Without a Model | Using a high Update-To-Data (UTD) ratio, model-based methods have recently achieved much higher sample efficiency than previous model-free methods for continuous-action DRL benchmarks. In this paper, we introduce a simple model-free algorithm, Randomized Ensembled Double Q-Learning (REDQ), and show that its performance is just as good as, if not better than, a state-of-the-art model-based algorithm for the MuJoCo benchmark. Moreover, REDQ can achieve this performance using fewer parameters than the model-based method, and with less wall-clock run time. REDQ has three carefully integrated ingredients which allow it to achieve its high performance: (i) a UTD ratio $\gg 1$; (ii) an ensemble of Q functions; (iii) in-target minimization across a random subset of Q functions from the ensemble. Through carefully designed experiments, we provide a detailed analysis of REDQ and related model-free algorithms. To our knowledge, REDQ is the first successful model-free DRL algorithm for continuous-action spaces using a UTD ratio $\gg 1$. | Accept (Poster) |
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/2021/Conference | CPT: Efficient Deep Neural Network Training via Cyclic Precision | Low-precision deep neural network (DNN) training has gained tremendous attention as reducing precision is one of the most effective knobs for boosting DNNs' training time/energy efficiency. In this paper, we attempt to explore low-precision training from a new perspective as inspired by recent findings in understanding DNN training: we conjecture that DNNs' precision might have a similar effect as the learning rate during DNN training, and advocate dynamic precision along the training trajectory for further boosting the time/energy efficiency of DNN training. Specifically, we propose Cyclic Precision Training (CPT) to cyclically vary the precision between two boundary values which can be identified using a simple precision range test within the first few training epochs. Extensive simulations and ablation studies on five datasets and eleven models demonstrate that CPT's effectiveness is consistent across various models/tasks (including classification and language modeling). Furthermore, through experiments and visualization we show that CPT helps to (1) converge to a wider minima with a lower generalization error and (2) reduce training variance which we believe opens up a new design knob for simultaneously improving the optimization and efficiency of DNN training. | Accept (Spotlight) |
ICLR.cc/2020/Conference | Global-Local Network for Learning Depth with Very Sparse Supervision | Natural intelligent agents learn to perceive the three dimensional structure of the world without training on large datasets and are unlikely to have the precise equations of projective geometry hard-wired in the brain. Such skill would also be valuable to artificial systems in order to avoid the expensive collection of labeled datasets, as well as tedious tuning required by methods based on multi-view geometry. Inspired by natural agents, who interact with the environment via visual and haptic feedback, this paper explores a new approach to learning depth from images and very sparse depth measurements, just a few pixels per image. To learn from such extremely sparse supervision, we introduce an appropriate inductive bias by designing a specialized global-local network architecture. Experiments on several datasets show that the proposed model can learn monocular dense depth estimation when trained with very sparse ground truth, even a single pixel per image. Moreover, we find that the global parameters extracted by the network are predictive of the metric agent motion. | Reject |
ICLR.cc/2022/Conference | Token Pooling in Vision Transformers | Despite the recent success in many applications, the high computational requirements of vision transformers limit their use in resource-constrained settings. While many existing methods improve the quadratic complexity of attention, in most vision transformers, self-attention is not the major computation bottleneck, e.g., more than 80% of the computation is spent on fully-connected layers. To improve the computational complexity of all layers, we propose a novel token downsampling method, called Token Pooling, efficiently exploiting redundancies in the images and intermediate token representations. We show that, under mild assumptions, softmax-attention acts as a high-dimensional low-pass (smoothing) filter. Thus, its output contains redundancy that can be pruned to achieve a better trade-off between the computational cost and accuracy. Our new technique accurately approximates a set of tokens by minimizing the reconstruction error caused by downsampling. We solve this optimization problem via cost-efficient clustering. We rigorously analyze and compare to prior downsampling methods. Our experiments show that Token Pooling significantly improves the cost-accuracy trade-off over the state-of-the-art downsampling. Token Pooling is a simple and effective operator that can benefit many architectures. Applied to DeiT, it achieves the same ImageNet top-1 accuracy using 42% fewer computations. | Reject |
ICLR.cc/2021/Conference | Unbiased Learning with State-Conditioned Rewards in Adversarial Imitation Learning | Adversarial imitation learning has emerged as a general and scalable framework for automatic reward acquisition. However, we point out that previous methods commonly exploited occupancy-dependent reward learning formulation—which hinders the reconstruction of optimal decision as an energy-based model. Despite the theoretical justification, the occupancy measures tend to cause issues in practice because of high variance and low vulnerability to domain shifts. Another reported problem is termination biases induced by provided rewarding and regularization schemes around terminal states. In order to deal with these issues, this work presents a novel algorithm called causal adversarial inverse reinforcement learning. Our formulation draws a strong connection between adversarial learning and energy-based reinforcement learning; thus, the architecture is capable of recovering a reward function that induces a multi-modal policy. In experiments, we demonstrate that our approach outperforms prior methods in challenging continuous control tasks, even under significant variation in the environments. | Reject |
ICLR.cc/2022/Conference | Stochastic Reweighted Gradient Descent | Importance sampling is a promising strategy for improving the convergence rate of stochastic gradient methods. It is typically used to precondition the optimization problem, but it can also be used to reduce the variance of the gradient estimator. Unfortunately, this latter point of view has yet to lead to practical methods that improve the asymptotic error of stochastic gradient methods. In this work, we propose stochastic reweighted gradient (SRG), a variance-reduced stochastic gradient method based solely on importance sampling that can improve on the asymptotic error of stochastic gradient descent (SGD) in the strongly convex and smooth case. We show that SRG can be extended to combine the benefits of both importance-sampling-based preconditioning and variance reduction. When compared to SGD, the resulting algorithm can simultaneously reduce the condition number and the asymptotic error, both by up to a factor equal to the number of component functions. We demonstrate improved convergence in practice on $\ell_2$-regularized logistic regression problems. | Reject |
ICLR.cc/2022/Conference | SimVLM: Simple Visual Language Model Pretraining with Weak Supervision | With recent progress in joint modeling of visual and textual representations, Vision-Language Pretraining (VLP) has achieved impressive performance on many multimodal downstream tasks. However, the requirement for expensive annotations including clean image captions and regional labels limits the scalability of existing approaches, and complicates the pretraining procedure with the introduction of multiple dataset-specific objectives. In this work, we relax these constraints and present a minimalist pretraining framework, named Simple Visual Language Model (SimVLM). Unlike prior work, SimVLM reduces the training complexity by exploiting large-scale weak supervision, and is trained end-to-end with a single prefix language modeling objective. Without utilizing extra data or task-specific customization, the resulting model significantly outperforms previous pretraining methods and achieves new state-of-the-art results on a wide range of discriminative and generative vision-language benchmarks, including VQA (+3.74% vqa-score), NLVR2 (+1.17% accuracy), SNLI-VE (+1.37% accuracy) and image captioning tasks (+10.1% average CIDEr score). Furthermore, we demonstrate that SimVLM acquires strong generalization and transfer ability, enabling zero-shot behavior including open-ended visual question answering and cross-modality transfer. | Accept (Poster) |
ICLR.cc/2022/Conference | Learning Fast, Learning Slow: A General Continual Learning Method based on Complementary Learning System | Humans excel at continually learning from an ever-changing environment whereas it remains a challenge for deep neural networks which exhibit catastrophic forgetting. The complementary learning system (CLS) theory suggests that the interplay between rapid instance-based learning and slow structured learning in the brain is crucial for accumulating and retaining knowledge. Here, we propose CLS-ER, a novel dual memory experience replay (ER) method which maintains short-term and long-term semantic memories that interact with the episodic memory. Our method employs an effective replay mechanism whereby new knowledge is acquired while aligning the decision boundaries with the semantic memories. CLS-ER does not utilize the task boundaries or make any assumption about the distribution of the data which makes it versatile and suited for ``general continual learning''. Our approach achieves state-of-the-art performance on standard benchmarks as well as more realistic general continual learning settings. | Accept (Poster) |
ICLR.cc/2020/Conference | MODELLING BIOLOGICAL ASSAYS WITH ADAPTIVE DEEP KERNEL LEARNING | Due to the significant costs of data generation, many prediction tasks within drug discovery are by nature few-shot regression (FSR) problems, including accurate modelling of biological assays. Although a number of few-shot classification and reinforcement learning methods exist for similar applications, we find relatively few FSR methods meeting the performance standards required for such tasks under real-world constraints. Inspired by deep kernel learning, we develop a novel FSR algorithm that is better suited to these settings. Our algorithm consists of learning a deep network in combination with a kernel function and a differentiable kernel algorithm. As the choice of the kernel is critical, our algorithm learns to find the appropriate one for each task during inference. It thus performs more effectively with complex task distributions, outperforming current state-of-the-art algorithms on both toy and novel, real-world benchmarks that we introduce herein. By introducing novel benchmarks derived from biological assays, we hope that the community will progress towards the development of FSR algorithms suitable for use in noisy and uncertain environments such as drug discovery. | Reject |
ICLR.cc/2018/Conference | Lifelong Word Embedding via Meta-Learning | Learning high-quality word embeddings is of significant importance in achieving better performance in many down-stream learning tasks. On one hand, traditional word embeddings are trained on a large scale corpus for general-purpose tasks, which are often sub-optimal for many domain-specific tasks. On the other hand, many domain-specific tasks do not have a large enough domain corpus to obtain high-quality embeddings. We observe that domains are not isolated and a small domain corpus can leverage the learned knowledge from many past domains to augment that corpus in order to generate high-quality embeddings. In this paper, we formulate the learning of word embeddings as a lifelong learning process. Given knowledge learned from many previous domains and a small new domain corpus, the proposed method can effectively generate new domain embeddings by leveraging a simple but effective algorithm and a meta-learner, where the meta-learner is able to provide word context similarity information at the domain-level. Experimental results demonstrate that the proposed method can effectively learn new domain embeddings from a small corpus and past domain knowledges\footnote{We will release the code after final revisions.}. We also demonstrate that general-purpose embeddings trained from a large scale corpus are sub-optimal in domain-specific tasks. | Reject |
ICLR.cc/2023/Conference | The KFIoU Loss for Rotated Object Detection | Differing from the well-developed horizontal object detection area whereby the computing-friendly IoU based loss is readily adopted and well fits with the detection metrics, rotation detectors often involve a more complicated loss based on SkewIoU which is unfriendly to gradient-based training. In this paper, we propose an effective approximate SkewIoU loss based on Gaussian modeling and Gaussian product, which mainly consists of two items. The first term is a scale-insensitive center point loss, which is used to quickly narrow the distance between the center points of the two bounding boxes. In the distance-independent second term, the product of the Gaussian distributions is adopted to inherently mimic the mechanism of SkewIoU by its definition, and show its alignment with the SkewIoU loss at trend-level within a certain distance (i.e. within 9 pixels). This is in contrast to recent Gaussian modeling based rotation detectors e.g. GWD loss and KLD loss that involve a human-specified distribution distance metric which require additional hyperparameter tuning that vary across datasets and detectors. The resulting new loss called KFIoU loss is easier to implement and works better compared with exact SkewIoU loss, thanks to its full differentiability and ability to handle the non-overlapping cases. We further extend our technique to the 3-D case which also suffers from the same issues as 2-D. Extensive results on various datasets with different base detectors show the effectiveness of our approach. | Accept: poster |
ICLR.cc/2022/Conference | Towards understanding how momentum improves generalization in deep learning | Stochastic gradient descent (SGD) with momentum is widely used for training modern deep learning architectures. While it is well understood that using momentum can lead to faster convergence rate in various settings, it has also been observed that momentum yields higher generalization. Prior work argue that momentum stabilizes the SGD noise during training and this leads to higher generalization. In this paper, we take the opposite view to this result and first empirically show that gradient descent with momentum (GD+M) significantly improves generalization comparing to gradient descent (GD) in many deep learning tasks. From this observation, we formally study how momentum improves generalization in deep learning. We devise a binary classification setting where a two-layer (over-parameterized) convolutional neural network trained with GD+M provably generalizes better than the same network trained with vanilla GD, when both algorithms start from the same random initialization. The key insight in our analysis is that momentum is beneficial in datasets where the examples share some features but differ in their margin. Contrary to the GD model that memorizes the small margin data, GD+M can still learn the features in these data thanks to its historical gradients. We also empirically verify this learning process of momentum in real-world settings. | Reject |
ICLR.cc/2019/Conference | Unsupervised Document Representation using Partition Word-Vectors Averaging | Learning effective document-level representation is essential in many important NLP tasks such as document classification, summarization, etc. Recent research has shown that simple weighted averaging of word vectors is an effective way to represent sentences, often outperforming complicated seq2seq neural models in many tasks. While it is desirable to use the same method to represent documents as well, unfortunately, the effectiveness is lost when representing long documents involving multiple sentences. One reason for this degradation is due to the fact that a longer document is likely to contain words from many different themes (or topics), and hence creating a single vector while ignoring all the thematic structure is unlikely to yield an effective representation of the document. This problem is less acute in single sentences and other short text fragments where presence of a single theme/topic is most likely. To overcome this problem, in this paper we present PSIF, a partitioned word averaging model to represent long documents. P-SIF retains the simplicity of simple weighted word averaging while taking a document's thematic structure into account. In particular, P-SIF learns topic-specific vectors from a document and finally concatenates them all to represent the overall document. Through our experiments over multiple real-world datasets and tasks, we demonstrate PSIF's effectiveness compared to simple weighted averaging and many other state-of-the-art baselines. We also show that PSIF is particularly effective in representing long multi-sentence documents. We will release PSIF's embedding source code and data-sets for reproducing results. | Reject |
ICLR.cc/2020/Conference | POP-Norm: A Theoretically Justified and More Accelerated Normalization Approach | Batch Normalization (BatchNorm) has been a default module in modern deep networks due to its effectiveness for accelerating training deep neural networks. It is widely accepted that the great success of BatchNorm is owing to reduction of internal covariate shift (ICS), but recently it is demonstrated that the link between them is fairly weak. The intrinsic reason behind effectiveness of BatchNorm is still unrevealed that limits it to be made better use. In light of this, we propose a new normalization approach, referred to as Pre-Operation Normalization (POP-Norm), which is theoretically ensured to speed up the training convergence. Not surprisingly, POP-Norm and BatchNorm are largely the same. Hence the similarities can help us to theoretically interpret the root of BatchNorm's effectiveness. There are still some significant distinctions between the two approaches. Just the distinctions make POP-Norm achieve faster convergence rate and better performance than BatchNorm, which are validated in extensive experiments on benchmark datasets: CIFAR10, CIFAR100 and ILSVRC2012. | Reject |
ICLR.cc/2023/Conference | CoRTX: Contrastive Framework for Real-time Explanation | Recent advancements in explainable machine learning provide effective and faithful solutions for interpreting model behaviors. However, many explanation methods encounter efficiency issues, which largely limit their deployments in practical scenarios. Real-time explainer (RTX) frameworks have thus been proposed to accelerate the model explanation process by learning an one-feed-forward explainer. Existing RTX frameworks typically build the explainer under the supervised learning paradigm, which requires large amounts of explanation labels as the ground truth. Considering that accurate explanation labels are usually hard to obtain, due to constrained computational resources and limited human efforts, effective explainer training is still challenging in practice. In this work, we propose a COntrastive Real-Time eXplanation (CoRTX) framework to learn the explanation-oriented representation and relieve the intensive dependence of explainer training on explanation labels. Specifically, we design a synthetic strategy to select positive and negative instances for explanation representation learning. Theoretical analysis show that our selection strategy can benefit the contrastive learning process on explanation tasks. Experimental results on three real-world datasets further demonstrate the efficiency and efficacy of our proposed CoRTX framework. | Accept: poster |
ICLR.cc/2021/Conference | Nonseparable Symplectic Neural Networks | Predicting the behaviors of Hamiltonian systems has been drawing increasing attention in scientific machine learning. However, the vast majority of the literature was focused on predicting separable Hamiltonian systems with their kinematic and potential energy terms being explicitly decoupled, while building data-driven paradigms to predict nonseparable Hamiltonian systems that are ubiquitous in fluid dynamics and quantum mechanics were rarely explored. The main computational challenge lies in the effective embedding of symplectic priors to describe the inherently coupled evolution of position and momentum, which typically exhibits intricate dynamics. To solve the problem, we propose a novel neural network architecture, Nonseparable Symplectic Neural Networks (NSSNNs), to uncover and embed the symplectic structure of a nonseparable Hamiltonian system from limited observation data. The enabling mechanics of our approach is an augmented symplectic time integrator to decouple the position and momentum energy terms and facilitate their evolution. We demonstrated the efficacy and versatility of our method by predicting a wide range of Hamiltonian systems, both separable and nonseparable, including chaotic vortical flows. We showed the unique computational merits of our approach to yield long-term, accurate, and robust predictions for large-scale Hamiltonian systems by rigorously enforcing symplectomorphism. | Accept (Poster) |
ICLR.cc/2018/Conference | APPLICATION OF DEEP CONVOLUTIONAL NEURAL NETWORK TO PREVENT ATM FRAUD BY FACIAL DISGUISE IDENTIFICATION | The paper proposes and demonstrates a Deep Convolutional Neural Network (DCNN) architecture to identify users with disguised face attempting a fraudulent ATM transaction. The recent introduction of Disguised Face Identification (DFI) framework proves the applicability of deep neural networks for this very problem. All the ATMs nowadays incorporate a hidden camera in them and capture the footage of their users. However, it is impossible for the police to track down the impersonators with disguised faces from the ATM footage. The proposed deep convolutional neural network is trained to identify, in real time, whether the user in the captured image is trying to cloak his identity or not. The output of the DCNN is then reported to the ATM to take appropriate steps and prevent the swindler from completing the transaction. The network is trained using a dataset of images captured in similar situations as of an ATM. The comparatively low background clutter in the images enables the network to demonstrate high accuracy in feature extraction and classification for all the different disguises. | Reject |
ICLR.cc/2021/Conference | Learning to live with Dale's principle: ANNs with separate excitatory and inhibitory units | The units in artificial neural networks (ANNs) can be thought of as abstractions of biological neurons, and ANNs are increasingly used in neuroscience research. However, there are many important differences between ANN units and real neurons. One of the most notable is the absence of Dale's principle, which ensures that biological neurons are either exclusively excitatory or inhibitory. Dale's principle is typically left out of ANNs because its inclusion impairs learning. This is problematic, because one of the great advantages of ANNs for neuroscience research is their ability to learn complicated, realistic tasks. Here, by taking inspiration from feedforward inhibitory interneurons in the brain we show that we can develop ANNs with separate populations of excitatory and inhibitory units that learn just as well as standard ANNs. We call these networks Dale's ANNs (DANNs). We present two insights that enable DANNs to learn well: (1) DANNs are related to normalization schemes, and can be initialized such that the inhibition centres and standardizes the excitatory activity, (2) updates to inhibitory neuron parameters should be scaled using corrections based on the Fisher Information matrix. These results demonstrate how ANNs that respect Dale's principle can be built without sacrificing learning performance, which is important for future work using ANNs as models of the brain. The results may also have interesting implications for how inhibitory plasticity in the real brain operates. | Accept (Poster) |
ICLR.cc/2020/Conference | Robust Learning with Jacobian Regularization | Design of reliable systems must guarantee stability against input perturbations. In machine learning, such guarantee entails preventing overfitting and ensuring robustness of models against corruption of input data. In order to maximize stability, we analyze and develop a computationally efficient implementation of Jacobian regularization that increases classification margins of neural networks. The stabilizing effect of the Jacobian regularizer leads to significant improvements in robustness, as measured against both random and adversarial input perturbations, without severely degrading generalization properties on clean data. | Reject |
ICLR.cc/2021/Conference | Interpretable Relational Representations for Food Ingredient Recommendation Systems | Supporting chefs with ingredient recommender systems to create new recipes is challenging, as good ingredient combinations depend on many factors like taste, smell, cuisine style, texture among others. There have been few attempts to address these issues using machine learning. Importantly, useful models do obviously need to be accurate but importantly -- especially for food professionals -- interpretable. In order to address these issues, we propose the Interpretable Relational Representation Model (IRRM). The main component of the model is a key-value memory network to represent relationships of ingredients. We propose and test two variants of the model.
One can learn latent relational representations over a trainable memory network (Implicit model), and the other can learn explainable relational representations over a pre-trained memory network that integrates an external knowledge base (Explicit model).
The relational representations resulting from the model are interpretable -- they allow to inspect why certain ingredient pairings have been suggested. The Explicit model additionally allows to integrate any number of manually specified constraints.
We conduct experiments on two recipe datasets, including CulinaryDB with 45,772 recipes and Flavornet with 55,001 recipes, respectively. The experimental results show that our models are both predictive and informative. | Reject |
ICLR.cc/2023/Conference | Quadratic models for understanding neural network dynamics | In this work, we show that recently proposed quadratic models capture optimization and generalization properties of wide neural networks that cannot be captured by linear models. In particular, we prove that quadratic models for shallow ReLU networks exhibit the "catapult phase" from Lewkowycz et al. (2020) that arises when training such models with large learning rates. We then empirically show that the behaviour of quadratic models parallels that of neural networks in generalization, especially in the catapult phase regime. Our analysis further demonstrates that quadratic models are an effective tool for analysis of neural networks. | Reject |
ICLR.cc/2021/Conference | Partitioned Learned Bloom Filters | Bloom filters are space-efficient probabilistic data structures that are used to test whether an element is a member of a set, and may return false positives. Recently, variations referred to as learned Bloom filters were developed that can provide improved performance in terms of the rate of false positives, by using a learned model for the represented set. However, previous methods for learned Bloom filters do not take full advantage of the learned model. Here we show how to frame the problem of optimal model utilization as an optimization problem, and using our framework derive algorithms that can achieve near-optimal performance in many cases. | Accept (Poster) |
ICLR.cc/2023/Conference | Sparse Mixture-of-Experts are Domain Generalizable Learners | Human visual perception can easily generalize to out-of-distributed visual data, which is far beyond the capability of modern machine learning models. Domain generalization (DG) aims to close this gap, with existing DG methods mainly focusing on the loss function design. In this paper, we propose to explore an orthogonal direction, i.e., the design of the backbone architecture. It is motivated by an empirical finding that transformer-based models trained with empirical risk minimization (ERM) outperform CNN-based models employing state-of-the-art (SOTA) DG algorithms on multiple DG datasets. We develop a formal framework to characterize a network's robustness to distribution shifts by studying its architecture's alignment with the correlations in the dataset. This analysis guides us to propose a novel DG model built upon vision transformers, namely \emph{Generalizable Mixture-of-Experts (GMoE)}. Extensive experiments on DomainBed demonstrate that GMoE trained with ERM outperforms SOTA DG baselines by a large margin. Moreover, GMoE is complementary to existing DG methods and its performance is substantially improved when trained with DG algorithms. | Accept: notable-top-5% |
ICLR.cc/2021/Conference | Deep Learning Is Composite Kernel Learning | Recent works have connected deep learning and kernel methods. In this paper, we show that architectural choices such as convolutional layers with pooling, skip connections, make deep learning a composite kernel learning method, where the kernel is a (architecture dependent) composition of base kernels: even before training, standard deep networks have in-built structural properties that ensure their success. In particular, we build on the recently developed `neural path' framework that characterises the role of gates/masks in fully connected deep networks with ReLU activations. | Reject |
ICLR.cc/2021/Conference | When Optimizing $f$-Divergence is Robust with Label Noise | We show when maximizing a properly defined $f$-divergence measure with respect to a classifier's predictions and the supervised labels is robust with label noise. Leveraging its variational form, we derive a nice decoupling property for a family of $f$-divergence measures when label noise presents, where the divergence is shown to be a linear combination of the variational difference defined on the clean distribution and a bias term introduced due to the noise. The above derivation helps us analyze the robustness of different $f$-divergence functions. With established robustness, this family of $f$-divergence functions arises as useful metrics for the problem of learning with noisy labels, which do not require the specification of the labels' noise rate. When they are possibly not robust, we propose fixes to make them so. In addition to the analytical results, we present thorough experimental evidence. Our code is available at https://github.com/UCSC-REAL/Robust-f-divergence-measures. | Accept (Poster) |