text
stringlengths 23
5.09k
| label
class label 3
classes |
|---|---|
Title: Variational oracle guiding for reinforcement learning. Abstract: How to make intelligent decisions is a central problem in machine learning and artificial intelligence. Despite recent successes of deep reinforcement learning (RL) in various decision making problems, an important but under-explored aspect is how to leverage oracle observation (the information that is invisible during online decision making, but is available during offline training) to facilitate learning. For example, human experts will look at the replay after a Poker game, in which they can check the opponents' hands to improve their estimation of the opponents' hands from the visible information during playing. In this work, we study such problems based on Bayesian theory and derive an objective to leverage oracle observation in RL using variational methods. Our key contribution is to propose a general learning framework referred to as variational latent oracle guiding (VLOG) for DRL. VLOG is featured with preferable properties such as its robust and promising performance and its versatility to incorporate with any value-based DRL algorithm. We empirically demonstrate the effectiveness of VLOG in online and offline RL domains with tasks ranging from video games to a challenging tile-based game Mahjong. Furthermore, we publish the Mahjong environment and an offline RL dataset as a benchmark to facilitate future research on oracle guiding (https://github.com/Agony5757/mahjong). | 1accept
|
Title: On the Certified Robustness for Ensemble Models and Beyond. Abstract: Recent studies show that deep neural networks (DNN) are vulnerable to adversarial examples, which aim to mislead DNNs by adding perturbations with small magnitude. To defend against such attacks, both empirical and theoretical defense approaches have been extensively studied for a single ML model. In this work, we aim to analyze and provide the certified robustness for ensemble ML models, together with the sufficient and necessary conditions of robustness for different ensemble protocols. Although ensemble models are shown more robust than a single model empirically; surprisingly, we find that in terms of the certified robustness the standard ensemble models only achieve marginal improvement compared to a single model. Thus, to explore the conditions that guarantee to provide certifiably robust ensemble ML models, we first prove that diversified gradient and large confidence margin are sufficient and necessary conditions for certifiably robust ensemble models under the model-smoothness assumption. We then provide the bounded model-smoothness analysis based on the proposed Ensemble-before-Smoothing strategy. We also prove that an ensemble model can always achieve higher certified robustness than a single base model under mild conditions. Inspired by the theoretical findings, we propose the lightweight Diversity Regularized Training (DRT) to train certifiably robust ensemble ML models. Extensive experiments show that our DRT enhanced ensembles can consistently achieve higher certified robustness than existing single and ensemble ML models, demonstrating the state-of-the-art certified $L_2$-robustness on MNIST, CIFAR-10, and ImageNet datasets. | 1accept
|
Title: Semi-Empirical Objective Functions for Neural MCMC Proposal Optimization. Abstract: Current objective functions used for training neural MCMC proposal distributions implicitly rely on architectural restrictions to yield sensible optimization results, which hampers the development of highly expressive neural MCMC proposal architectures. In this work, we introduce and demonstrate a semi-empirical procedure for determining approximate objective functions suitable for optimizing arbitrarily parameterized proposal distributions in MCMC methods. Our proposed Ab Initio objective functions consist of the weighted combination of functions following constraints on their global optima and transformation invariances that we argue should be upheld by general measures of MCMC efficiency for use in proposal optimization. Our experimental results demonstrate that Ab Initio objective functions maintain favorable performance and preferable optimization behavior compared to existing objective functions for neural MCMC optimization. We find that Ab Initio objective functions are sufficiently robust to enable the confident optimization of neural proposal distributions parameterized by deep generative networks extending beyond the regimes of traditional MCMC schemes. | 0reject
|
Title: AMRL: Aggregated Memory For Reinforcement Learning. Abstract: In many partially observable scenarios, Reinforcement Learning (RL) agents must rely on long-term memory in order to learn an optimal policy. We demonstrate that using techniques from NLP and supervised learning fails at RL tasks due to stochasticity from the environment and from exploration. Utilizing our insights on the limitations of traditional memory methods in RL, we propose AMRL, a class of models that can learn better policies with greater sample efficiency and are resilient to noisy inputs. Specifically, our models use a standard memory module to summarize short-term context, and then aggregate all prior states from the standard model without respect to order. We show that this provides advantages both in terms of gradient decay and signal-to-noise ratio over time. Evaluating in Minecraft and maze environments that test long-term memory, we find that our model improves average return by 19% over a baseline that has the same number of parameters and by 9% over a stronger baseline that has far more parameters. | 1accept
|
Title: Deep Q-Network with Proximal Iteration. Abstract: We employ Proximal Iteration for value-function optimization in reinforcement learning. Proximal Iteration is a computationally efficient technique that enables us to bias the optimization procedure towards more desirable solutions. As a concrete application of Proximal Iteration in deep reinforcement learning, we endow the objective function of the Deep Q-Network (DQN) agent with a proximal term to ensure that the online-network component of DQN remains in the vicinity of the target network. The resultant agent, which we call DQN with Proximal Iteration, or DQNPro, exhibits significant improvements over the original DQN on the Atari benchmark. Our results accentuate the power of employing sound optimization techniques for deep reinforcement learning. | 0reject
|
Title: STRATA: Simple, Gradient-free Attacks for Models of Code. Abstract: Adversarial examples are imperceptible perturbations in the input to a neural model that result in misclassification. Generating adversarial examples for source code poses an additional challenge compared to the domains of images and natural language, because source code perturbations must adhere to strict semantic guidelines so the resulting programs retain the functional meaning of the code. We propose a simple and efficient gradient-free method for generating state-of-the-art adversarial examples on models of code that can be applied in a white-box or black-box setting. Our method generates untargeted and targeted attacks, and empirically outperforms competing gradient-based methods with less information and less computational effort. | 0reject
|
Title: Warpspeed Computation of Optimal Transport, Graph Distances, and Embedding Alignment. Abstract: Optimal transport (OT) is a cornerstone of many machine learning tasks. The current best practice for computing OT is via entropy regularization and Sinkhorn iterations. This algorithm runs in quadratic time and requires calculating the full pairwise cost matrix, which is prohibitively expensive for large sets of objects. To alleviate this limitation we propose to instead use a sparse approximation of the cost matrix based on locality sensitive hashing (LSH). Moreover, we fuse this sparse approximation with the Nyström method, resulting in the locally corrected Nyström method (LCN). These approximations enable general log-linear time algorithms for entropy-regularized OT that perform well even in complex, high-dimensional spaces. We thoroughly demonstrate these advantages via a theoretical analysis and by evaluating multiple approximations both directly and as a component of two real-world models. Using approximate Sinkhorn for unsupervised word embedding alignment enables us to train the model full-batch in a fraction of the time while improving upon the original on average by 3.1 percentage points without any model changes. For graph distance regression we propose the graph transport network (GTN), which combines graph neural networks (GNNs) with enhanced Sinkhorn and outcompetes previous models by 48%. LCN-Sinkhorn enables GTN to achieve this while still scaling log-linearly in the number of nodes. | 0reject
|
Title: AA-PINN: ATTENTION AUGMENTED PHYSICS INFORMED NEURAL NETWORKS. Abstract: Physics Informed Neural Networks has been quite successful in modelling the complex nature of fluid flow. Computational Fluid Dynamics using parallel processing
algorithms on GPUs have considerably reduced the time to solve the Navier Stokes
Equations. CFD based approaches uses approximates to make the modelling easy
but it comes at the cost of decrease in accuracy. In this paper, we propose an
attention based network architecture named AA-PINN to model PDEs behind fluid
flow. We use a combination of channel and spatial attention module. We propose a
novel loss function which is more robust in handling the initial as well as boundary
conditions imposed. Using evaluation metrics like RMSE, divergence and thermal
kinetic energy, our network outperforms previous PINNs for modelling Navier
Stokes and Burgers Equation. | 0reject
|
Title: Asynchronous SGD without gradient delay for efficient distributed training. Abstract: Asynchronous distributed gradient descent algorithms for training of deep neural
networks are usually considered as inefficient, mainly because of the Gradient delay
problem. In this paper, we propose a novel asynchronous distributed algorithm
that tackles this limitation by well-thought-out averaging of model updates, computed
by workers. The algorithm allows computing gradients along the process
of gradient merge, thus, reducing or even completely eliminating worker idle time
due to communication overhead, which is a pitfall of existing asynchronous methods.
We provide theoretical analysis of the proposed asynchronous algorithm,
and show its regret bounds. According to our analysis, the crucial parameter for
keeping high convergence rate is the maximal discrepancy between local parameter
vectors of any pair of workers. As long as it is kept relatively small, the
convergence rate of the algorithm is shown to be the same as the one of a sequential
online learning. Furthermore, in our algorithm, this discrepancy is bounded
by an expression that involves the staleness parameter of the algorithm, and is
independent on the number of workers. This is the main differentiator between
our approach and other solutions, such as Elastic Asynchronous SGD or Downpour
SGD, in which that maximal discrepancy is bounded by an expression that
depends on the number of workers, due to gradient delay problem. To demonstrate
effectiveness of our approach, we conduct a series of experiments on image
classification task on a cluster with 4 machines, equipped with a commodity communication
switch and with a single GPU card per machine. Our experiments
show a linear scaling on 4-machine cluster without sacrificing the test accuracy,
while eliminating almost completely worker idle time. Since our method allows
using commodity communication switch, it paves a way for large scale distributed
training performed on commodity clusters. | 0reject
|
Title: Meta-Learning Deep Energy-Based Memory Models. Abstract: We study the problem of learning an associative memory model -- a system which is able to retrieve a remembered pattern based on its distorted or incomplete version.
Attractor networks provide a sound model of associative memory: patterns are stored as attractors of the network dynamics and associative retrieval is performed by running the dynamics starting from a query pattern until it converges to an attractor.
In such models the dynamics are often implemented as an optimization procedure that minimizes an energy function, such as in the classical Hopfield network.
In general it is difficult to derive a writing rule for a given dynamics and energy that is both compressive and fast.
Thus, most research in energy-based memory has been limited either to tractable energy models not expressive enough to handle complex high-dimensional objects such as natural images, or to models that do not offer fast writing.
We present a novel meta-learning approach to energy-based memory models (EBMM) that allows one to use an arbitrary neural architecture as an energy model and quickly store patterns in its weights.
We demonstrate experimentally that our EBMM approach can build compressed memories for synthetic and natural data, and is capable of associative retrieval that outperforms existing memory systems in terms of the reconstruction error and compression rate. | 1accept
|
Title: TimeVAE: A Variational Auto-Encoder for Multivariate Time Series Generation. Abstract: Recent work in synthetic data generation in the time-series domain has focused on the use of Generative Adversarial Networks. We propose a novel architecture for synthetically generating time-series data with the use of Variational Auto-Encoders (VAEs). The proposed architecture has several distinct properties: interpretability, ability to encode domain knowledge, and reduced training times. We evaluate data generation quality by similarity and predictability against four multivariate datasets. We experiment with varying sizes of training data to measure the impact of data availability on generation quality for our VAE method as well as several state-of-the-art data generation methods. Our results on similarity tests show that the VAE approach is able to accurately represent the temporal attributes of the original data. On next-step prediction tasks using generated data, the proposed VAE architecture consistently meets or exceeds performance of state-of-the-art data generation methods. While noise reduction may cause the generated data to deviate from original data, we demonstrate the resulting de-noised data can significantly improve performance for next-step prediction using generated data. Finally, the proposed architecture can incorporate domain-specific time-patterns such as polynomial trends and seasonalities to provide interpretable outputs. Such interpretability can be highly advantageous in applications requiring transparency of model outputs or where users desire to inject prior knowledge of time-series patterns into the generative model. | 0reject
|
Title: QTN-VQC: An End-to-End Learning Framework for Quantum Neural Networks. Abstract: The advent of noisy intermediate-scale quantum (NISQ) computers raises a crucial challenge to design quantum neural networks for fully quantum learning tasks. To bridge the gap, this work proposes an end-to-end learning framework named QTN-VQC, by introducing a trainable quantum tensor network (QTN) for quantum embedding on a variational quantum circuit (VQC). The architecture of QTN is composed of a parametric tensor-train network for feature extraction and a tensor product encoding for quantum encoding. We highlight the QTN for quantum embedding in terms of two perspectives: (1) we theoretically characterize QTN by analyzing its representation power of input features; (2) QTN enables an end-to-end parametric model pipeline, namely QTN-VQC, from the generation of quantum embedding to the output measurement. Our experiments on the MNIST dataset demonstrate the advantages of QTN for quantum embedding over other quantum embedding approaches. | 0reject
|
Title: Semi-supervised Offline Reinforcement Learning with Pre-trained Decision Transformers. Abstract: Pre-training deep neural network models using large unlabelled datasets followed by fine-tuning them on small task-specific datasets has emerged as a dominant paradigm in natural language processing (NLP) and computer vision (CV). Despite the widespread success, such a paradigm has remained atypical in reinforcement learning (RL).
In this paper, we investigate how we can leverage large reward-free (i.e. task-agnostic) offline datasets of prior interactions to pre-train agents that can then be fine-tuned using a small reward-annotated dataset. To this end, we present Pre-trained Decision Transformer (PDT), a simple yet powerful algorithm for semi-supervised Offline RL. By masking reward tokens during pre-training, the transformer learns to autoregressivley predict actions based on previous state and action context and effectively extracts behaviors present in the dataset. During fine-tuning, rewards are un-masked and the agent learns the set of skills that should be invoked for the desired behavior as per the reward function. We demonstrate the efficacy of this simple and flexible approach on tasks from the D4RL benchmark with limited reward annotations. | 0reject
|
Title: Improving Adversarial Robustness Requires Revisiting Misclassified Examples. Abstract: Deep neural networks (DNNs) are vulnerable to adversarial examples crafted by imperceptible perturbations. A range of defense techniques have been proposed to improve DNN robustness to adversarial examples, among which adversarial training has been demonstrated to be the most effective. Adversarial training is often formulated as a min-max optimization problem, with the inner maximization for generating adversarial examples. However, there exists a simple, yet easily overlooked fact that adversarial examples are only defined on correctly classified (natural) examples, but inevitably, some (natural) examples will be misclassified during training. In this paper, we investigate the distinctive influence of misclassified and correctly classified examples on the final robustness of adversarial training. Specifically, we find that misclassified examples indeed have a significant impact on the final robustness. More surprisingly, we find that different maximization techniques on misclassified examples may have a negligible influence on the final robustness, while different minimization techniques are crucial. Motivated by the above discovery, we propose a new defense algorithm called {\em Misclassification Aware adveRsarial Training} (MART), which explicitly differentiates the misclassified and correctly classified examples during the training. We also propose a semi-supervised extension of MART, which can leverage the unlabeled data to further improve the robustness. Experimental results show that MART and its variant could significantly improve the state-of-the-art adversarial robustness. | 1accept
|
Title: Learning Hyperbolic Representations of Topological Features. Abstract: Learning task-specific representations of persistence diagrams is an important problem in topological data analysis and machine learning. However, current state of the art methods are restricted in terms of their expressivity as they are focused on Euclidean representations. Persistence diagrams often contain features of infinite persistence (i.e., essential features) and Euclidean spaces shrink their importance relative to non-essential features because they cannot assign infinite distance to finite points. To deal with this issue, we propose a method to learn representations of persistence diagrams on hyperbolic spaces, more specifically on the Poincare ball. By representing features of infinite persistence infinitesimally close to the boundary of the ball, their distance to non-essential features approaches infinity, thereby their relative importance is preserved. This is achieved without utilizing extremely high values for the learnable parameters, thus the representation can be fed into downstream optimization methods and trained efficiently in an end-to-end fashion. We present experimental results on graph and image classification tasks and show that the performance of our method is on par with or exceeds the performance of other state of the art methods.
| 1accept
|
Title: Is Homophily a Necessity for Graph Neural Networks?. Abstract: Graph neural networks (GNNs) have shown great prowess in learning representations suitable for numerous graph-based machine learning tasks. When applied to semi-supervised node classification, GNNs are widely believed to work well due to the homophily assumption (``like attracts like''), and fail to generalize to heterophilous graphs where dissimilar nodes connect. Recent works design new architectures to overcome such heterophily-related limitations, citing poor baseline performance and new architecture improvements on a few heterophilous graph benchmark datasets as evidence for this notion. In our experiments, we empirically find that standard graph convolutional networks (GCNs) can actually achieve better performance than such carefully designed methods on some commonly used heterophilous graphs. This motivates us to reconsider whether homophily is truly necessary for good GNN performance. We find that this claim is not quite true, and in fact, GCNs can achieve strong performance on heterophilous graphs under certain conditions. Our work carefully characterizes these conditions and provides supporting theoretical understanding and empirical observations. Finally, we examine existing heterophilous graphs benchmarks and reconcile how the GCN (under)performs on them based on this understanding. | 1accept
|
Title: Leave no Trace: Learning to Reset for Safe and Autonomous Reinforcement Learning. Abstract: Deep reinforcement learning algorithms can learn complex behavioral skills, but real-world application of these methods requires a considerable amount of experience to be collected by the agent. In practical settings, such as robotics, this involves repeatedly attempting a task, resetting the environment between each attempt. However, not all tasks are easily or automatically reversible. In practice, this learning process requires considerable human intervention. In this work, we propose an autonomous method for safe and efficient reinforcement learning that simultaneously learns a forward and backward policy, with the backward policy resetting the environment for a subsequent attempt. By learning a value function for the backward policy, we can automatically determine when the forward policy is about to enter a non-reversible state, providing for uncertainty-aware safety aborts. Our experiments illustrate that proper use of the backward policy can greatly reduce the number of manual resets required to learn a task and can reduce the number of unsafe actions that lead to non-reversible states. | 1accept
|
Title: The Deep Generative Decoder: using MAP estimates of representations. Abstract: A deep generative model is characterized by a representation space, its distribution, and a neural network mapping the representation to a distribution over vectors in feature space. Common methods such as variational autoencoders (VAEs) apply variational inference for training the neural network, but optimizing these models is often non-trivial. The encoder adds to the complexity of the model and introduces an amortization gap and the quality of the variational approximation is usually unknown. Additionally, the balance of the loss terms of the objective function heavily influences performance. Therefore, we argue that it is worthwhile to investigate a much simpler approximation which finds representations and their distribution by maximizing the model likelihood via back-propagation. In this approach, there is no encoder, and we therefore call it a Deep Generative Decoder (DGD). Using the CIFAR10 data set, we show that the DGD is easier and faster to optimize than the VAE, achieves more consistent low reconstruction errors of test data, and alleviates the problem of balancing the reconstruction and distribution loss terms. Although the model in its simple form cannot compete with state-of-the-art image generation approaches, it obtains better image generation scores than the variational approach on the CIFAR10 data. We demonstrate on MNIST data how the use of a Gaussian mixture with priors can lead to a clear separation of classes in a 2D representation space, and how the DGD can be used with labels to obtain a supervised representation. | 2withdrawn
|
Title: Unsupervised Clustering using Pseudo-semi-supervised Learning. Abstract: In this paper, we propose a framework that leverages semi-supervised models to improve unsupervised clustering performance. To leverage semi-supervised models, we first need to automatically generate labels, called pseudo-labels. We find that prior approaches for generating pseudo-labels hurt clustering performance because of their low accuracy. Instead, we use an ensemble of deep networks to construct a similarity graph, from which we extract high accuracy pseudo-labels. The approach of finding high quality pseudo-labels using ensembles and training the semi-supervised model is iterated, yielding continued improvement. We show that our approach outperforms state of the art clustering results for multiple image and text datasets. For example, we achieve 54.6% accuracy for CIFAR-10 and 43.9% for 20news, outperforming state of the art by 8-12% in absolute terms. | 1accept
|
Title: Learning Weakly-supervised Contrastive Representations. Abstract: We argue that a form of the valuable information provided by the auxiliary information is its implied data clustering information. For instance, considering hashtags as auxiliary information, we can hypothesize that an Instagram image will be semantically more similar with the same hashtags. With this intuition, we present a two-stage weakly-supervised contrastive learning approach. The first stage is to cluster data according to its auxiliary information. The second stage is to learn similar representations within the same cluster and dissimilar representations for data from different clusters. Our empirical experiments suggest the following three contributions. First, compared to conventional self-supervised representations, the auxiliary-information-infused representations bring the performance closer to the supervised representations, which use direct downstream labels as supervision signals. Second, our approach performs the best in most cases, when comparing our approach with other baseline representation learning methods that also leverage auxiliary data information. Third, we show that our approach also works well with unsupervised constructed clusters (e.g., no auxiliary information), resulting in a strong unsupervised representation learning approach. | 1accept
|
Title: Constrained Physical-Statistics Models for Dynamical System Identification and Prediction. Abstract: Modeling dynamical systems combining prior physical knowledge and machine learning (ML) is promising in scientific problems when the underlying processes are not fully understood, e.g. when the dynamics is partially known. A common practice to identify the respective parameters of the physical and ML components is to formulate the problem as supervised learning on observed trajectories. However, this formulation leads to an infinite number of possible decompositions. To solve this ill-posedness, we reformulate the learning problem by introducing an upper bound on the prediction error of a physical-statistical model. This allows us to control the contribution of both the physical and statistical components to the overall prediction. This framework generalizes several existing hybrid schemes proposed in the literature. We provide theoretical guarantees on the well-posedness of our formulation along with a proof of convergence in a simple affine setting. For more complex dynamics, we validate our framework experimentally. | 1accept
|
Title: Demystifying Limited Adversarial Transferability in Automatic Speech Recognition Systems. Abstract: The targeted transferability of adversarial samples enables attackers to exploit black-box models in the real-world. The most popular method to produce these adversarial samples is optimization attacks, which have been shown to achieve a high level of transferability in some domains. However, recent research has demonstrated that these attack samples fail to transfer when applied to Automatic Speech Recognition Systems (ASRs). In this paper, we investigate factors preventing this transferability via exhaustive experimentation. To do so, we perform an ablation study on each stage of the ASR pipeline. We discover and quantify six factors (i.e., input type, MFCC, RNN, output type, and vocabulary and sequence sizes) that impact the targeted transferability of optimization attacks against ASRs. Future research can leverage our findings to build ASRs that are more robust to other transferable attack types (e.g., signal processing attacks), or to modify architectures in other domains to reduce their exposure to targeted transferability of optimization attacks. | 1accept
|
Title: Deformable Capsules for Object Detection. Abstract: Capsule networks promise significant benefits over convolutional networks by storing stronger internal representations, and routing information based on the agreement between intermediate representations' projections. Despite this, their success has been mostly limited to small-scale classification datasets due to their computationally expensive nature. Recent studies have partially overcome this burden by locally-constraining the dynamic routing of features with convolutional capsules. Though memory efficient, convolutional capsules impose geometric constraints which fundamentally limit the ability of capsules to model the pose/deformation of objects. Further, they do not address the bigger memory concern of class-capsules scaling-up to bigger tasks such as detection or large-scale classification. In this study, we introduce deformable capsules (DeformCaps), a new capsule structure (SplitCaps), and a novel dynamic routing algorithm (SE-Routing) to balance computational efficiency with the need for modeling a large number of objects and classes. We demonstrate that the proposed methods allow capsules to efficiently scale-up to large-scale computer vision tasks for the first time, and create the first-ever capsule network for object detection in the literature. Our proposed architecture is a one-stage detection framework and obtains results on MS COCO which are on-par with state-of-the-art one-stage CNN-based methods, while producing fewer false positive detections. | 0reject
|
Title: Deep Ensembles for Low-Data Transfer Learning. Abstract: In the low-data regime, it is difficult to train good supervised models from scratch.
Instead practitioners turn to pre-trained models, leveraging transfer learning. Ensembling is an empirically and theoretically appealing way to construct powerful predictive models, but the predominant approach of training multiple deep networks with different random initialisations collides with the need for transfer via pre-trained weights. In this work, we study different ways of creating ensembles from pre-trained models. We show that the nature of pre-training itself is a performant source of diversity, and propose a practical algorithm that efficiently identifies a subset of pre-trained models for any downstream dataset. The approach is simple: Use nearest-neighbour accuracy to rank pre-trained models, fine-tune the best ones with a small hyperparameter sweep, and greedily construct an ensemble to minimise validation cross-entropy. When evaluated together with strong baselines on 19 different downstream tasks (the Visual Task Adaptation Benchmark), this achieves state-of-the-art performance at a much lower inference budget, even when selecting from over 2,000 pre-trained models. We also assess our ensembles on ImageNet variants and show improved robustness to distribution shift. | 0reject
|
Title: Multi-modal Self-Supervision from Generalized Data Transformations. Abstract: In the image domain, excellent representation can be learned by inducing invariance to content-preserving transformations, such as image distortions. In this paper, we show that, for videos, the answer is more complex, and that better results can be obtained by accounting for the interplay between invariance, distinctiveness, multiple modalities and time. We introduce Generalized Data Transformations (GDTs) as a way to capture this interplay. GDTs reduce most previous self-supervised approaches to a choice of data transformations, even when this was not the case in the original formulations. They also allow to choose whether the representation should be invariant or distinctive w.r.t. each effect and tell which combinations are valid, thus allowing us to explore the space of combinations systematically. We show in this manner that being invariant to certain transformations and distinctive to others is critical to learning effective video representations, improving the state-of-the-art by a large margin, and even surpassing supervised pretraining. We demonstrate results on a variety of downstream video and audio classification and retrieval tasks, on datasets such as HMDB-51, UCF-101, DCASE2014, ESC-50 and VGG-Sound. In particular, we achieve new state-of-the-art accuracies of 72.8% on HMDB-51 and 95.2% on UCF-101. | 0reject
|
Title: POLAR: A Polynomial Arithmetic Framework for Verifying Neural-Network Controlled Systems. Abstract: We propose POLAR, a \textbf{pol}ynomial \textbf{ar}ithmetic framework that leverages polynomial overapproximations with interval remainders for bounded-time reachability analysis of neural network-controlled systems (NNCSs).
Compared with existing arithmetic approaches that use standard Taylor models, our framework uses a novel approach to iteratively overapproximate the neuron output ranges layer-by-layer via a combination of Bernstein polynomial interpolation for continuous activation functions and Taylor model arithmetic for the other operations. This approach overcomes the main drawback in the standard Taylor model arithmetic, i.e. its inability to handle functions that cannot be well approximated by Taylor polynomials, and significantly improve the accuracy and efficiency of reachable states computation for NNCSs. To further tighten the overapproximation, our method keeps the Taylor model remainders symbolic under the linear mappings when propagating Taylor models across the neural-network controller.
We show that POLAR can be seamlessly integrated with existing Taylor model flowpipe construction techniques, and POLAR significantly outperforms the current state-of-the-art techniques on a suite of benchmarks. | 2withdrawn
|
Title: Equivalence of State Equations from Different Methods in High-dimensional Regression. Abstract: State equations were firstly introduced in the approximate message passing (AMP) to describe the mean square error (MSE) in compressed sensing. Since then a set of state equations have appeared in studies of logistic regression, robust estimator and other high-dimensional statistics problems. Recently, a convex Gaussian min-max theorem(CGMT) approach was proposed to study high-dimensional statistic problems accompanying with another set of different state equations. This Paper provides a uniform viewpoint on these methods and shows the equivalence of their reduction forms, which causes that the resulting SE are essentially equivalent and can be converted into the same expression through parameter transformations. Combining these results, we show that these different state equations are derived from several equivalent reduction forms. We believe this equivalence shed light on discovering a deeper structure in high dimensional statistics. | 0reject
|
Title: On the Spectral Bias of Neural Networks. Abstract: Neural networks are known to be a class of highly expressive functions able to fit even random input-output mappings with 100% accuracy. In this work we present properties of neural networks that complement this aspect of expressivity. By using tools from Fourier analysis, we show that deep ReLU networks are biased towards low frequency functions, meaning that they cannot have local fluctuations without affecting their global behavior. Intuitively, this property is in line with the observation that over-parameterized networks find simple patterns that generalize across data samples. We also investigate how the shape of the data manifold affects expressivity by showing evidence that learning high frequencies gets easier with increasing manifold complexity, and present a theoretical understanding of this behavior. Finally, we study the robustness of the frequency components with respect to parameter perturbation, to develop the intuition that the parameters must be finely tuned to express high frequency functions. | 0reject
|
Title: Fortified Networks: Improving the Robustness of Deep Networks by Modeling the Manifold of Hidden Representations. Abstract: Deep networks have achieved impressive results across a variety of important tasks. However, a known weakness is a failure to perform well when evaluated on data which differ from the training distribution, even if these differences are very small, as is the case with adversarial examples. We propose \emph{Fortified Networks}, a simple transformation of existing networks, which “fortifies” the hidden layers in a deep network by identifying when the hidden states are off of the data manifold, and maps these hidden states back to parts of the data manifold where the network performs well. Our principal contribution is to show that fortifying these hidden states improves the robustness of deep networks and our experiments (i) demonstrate improved robustness to standard adversarial attacks in both black-box and white-box threat models; (ii) suggest that our improvements are not primarily due to the problem of deceptively good results due to degraded quality in the gradient signal (the gradient masking problem) and (iii) show the advantage of doing this fortification in the hidden layers instead of the input space. We demonstrate improvements in adversarial robustness on three datasets (MNIST, Fashion MNIST, CIFAR10), across several attack parameters, both white-box and black-box settings, and the most widely studied attacks (FGSM, PGD, Carlini-Wagner). We show that these improvements are achieved across a wide variety of hyperparameters. | 0reject
|
Title: LATENT OPTIMIZATION VARIATIONAL AUTOENCODER FOR CONDITIONAL MOLECULAR GENERATION. Abstract: Variational autoencoder (VAE) is a generation algorithm, consisting of an encoder and a decoder, and the latent variable from the encoder is used as the input of the decoder.
VAE is widely used for image, audio and text generation tasks. In general, the training of VAE is at risk of posterior collapsing especially for long sequential data. To alleviate this, modified evidence lower bounds (ELBOs) were proposed. However, these approaches heuristically control training loss using a hyper-parameter, and it is not way to solve the fundamental problem of vanilla VAE.
In this paper, we propose a method to insert an optimization step of the latent variable and alternately update the encoder and decoder of conditional VAE for maximizing ELBOs.
In experiments, we applied the latent optimization VAE (LOVAE) on ZINC database, consisting of string representation of molecules, for the inverse molecular design.
We showed that the proposed LOVAE achieves better performance than vanilla VAE in terms of ELBOs and molecular generation performance. In addition, the proposed method showed better performance in property satisfaction and property maximization tasks compared to existing works. | 0reject
|
Title: A Large-scale Study on Training Sample Memorization in Generative Modeling. Abstract: Many recent developments on generative models for natural images have relied on heuristically-motivated metrics that can be easily gamed by memorizing a small sample from the true distribution or training a model directly to improve the metric.
In this work, we critically evaluate the gameability of the benchmarking procedure by running a competition which ultimately resulted in participants attempting to cheat. Our competition received over 11000 submitted models which allowed us to investigate memorization-aware metrics for measuring generative model performance. Specifically, we propose the Memorization-Informed Frechet Inception Distance (MiFID) and discuss ways to ensure that winning submissions were based on genuine improvements in perceptual quality. We evaluate the effectiveness of our benchmark by manually inspecting the code for the 1000 top-performing models and labeling different forms of memorization that were intentionally or unintentionally used. To facilitate future work on benchmarking generative models, we release generated images and our labels for these models as well as code to compute the MiFID metric. | 0reject
|
Title: DIVA: Domain Invariant Variational Autoencoder. Abstract: We consider the problem of domain generalization, namely, how to learn representations given data from a set of domains that generalize to data from a previously unseen domain. We propose the Domain Invariant Variational Autoencoder (DIVA), a generative model that tackles this problem by learning three independent latent subspaces, one for the domain, one for the class, and one for any residual variations. We highlight that due to the generative nature of our model we can also incorporate unlabeled data from known or previously unseen domains. To the best of our knowledge this has not been done before in a domain generalization setting. This property is highly desirable in fields like medical imaging where labeled data is scarce. We experimentally evaluate our model on the rotated MNIST benchmark and a malaria cell images dataset where we show that (i) the learned subspaces are indeed complementary to each other, (ii) we improve upon recent works on this task and (iii) incorporating unlabelled data can boost the performance even further. | 0reject
|
Title: CaptainGAN: Navigate Through Embedding Space For Better Text Generation. Abstract: Score-function-based text generation approaches such as REINFORCE, in general, suffer from high computational complexity and training instability problems. This is mainly due to the non-differentiable nature of the discrete space sampling and thus these methods have to treat the discriminator as a reward function and ignore the gradient information. In this paper, we propose a novel approach, CaptainGAN, which adopts the straight-through gradient estimator and introduces a ”re-centered” gradient estimation technique to steer the generator toward better text tokens through the embedding space. Our method is stable to train and converges quickly without maximum likelihood pre-training. On multiple metrics of text quality and diversity, our method outperforms existing GAN-based methods on natural language generation. | 0reject
|
Title: Streamlining EM into Auto-Encoder Networks. Abstract: We present a new deep neural network architecture, named EDGaM, for deep clustering. This architecture can seamlessly learn deep auto-encoders and capture common group features of complex inputs in the encoded latent space. The key idea is to introduce a differentiable Gaussian mixture neural network between an encoder and a decoder. In particular, EDGaM streamlines the iterative Expectation-Maximum (EM) algorithm of the Gaussian mixture models into network design and replaces the alternative update with a forward-backward optimization. Being differentiable, both network weights and clustering centroids in EDGaM can be learned simultaneously in an end-to-end manner through standard stochastic gradient descent. To avoid preserving too many sample-specific details, we use both the clustering centroid and the original latent embedding for decoding. Meanwhile, we distill the soft clustering assignment for each sample via entropy minimization such that a clear cluster structure is exhibited. Our experiments show that our method outperforms state-of-the-art unsupervised clustering techniques in terms of both efficiency and clustering performance. | 2withdrawn
|
Title: S2VG: Soft Stochastic Value Gradient method. Abstract: Model-based reinforcement learning (MBRL) has shown its advantages in sample-efficiency over model-free reinforcement learning (MFRL). Despite the impressive results it achieves, it still faces a trade-off between the ease of data generation and model bias. In this paper, we propose a simple and elegant model-based reinforcement learning algorithm called soft stochastic value gradient method (S2VG). S2VG combines the merits of the maximum-entropy reinforcement learning and MBRL, and exploits both real and imaginary data. In particular, we embed the model in the policy training and learn $Q$ and $V$ functions from the real (or imaginary) data set. Such embedding enables us to compute an analytic policy gradient through the back-propagation rather than the likelihood-ratio estimation, which can reduce the variance of the gradient estimation. We name our algorithm Soft Stochastic Value Gradient method to indicate its connection with the well-known stochastic value gradient method in \citep{heess2015Learning}. | 0reject
|
Title: CyCADA: Cycle-Consistent Adversarial Domain Adaptation. Abstract: Domain adaptation is critical for success in new, unseen environments.
Adversarial adaptation models applied in feature spaces discover domain invariant representations, but are difficult to visualize and sometimes fail to capture pixel-level and low-level domain shifts.
Recent work has shown that generative adversarial networks combined with cycle-consistency constraints are surprisingly effective at mapping images between domains, even without the use of aligned image pairs.
We propose a novel discriminatively-trained Cycle-Consistent Adversarial Domain Adaptation model.
CyCADA adapts representations at both the pixel-level and feature-level, enforces cycle-consistency while leveraging a task loss, and does not require aligned pairs. Our model can be applied in a variety of visual recognition and prediction settings.
We show new state-of-the-art results across multiple adaptation tasks, including digit classification and semantic segmentation of road scenes demonstrating transfer from synthetic to real world domains. | 0reject
|
Title: Curriculum learning as a tool to uncover learning principles in the brain . Abstract: We present a novel approach to use curricula to identify principles by which a system learns. Previous work in curriculum learning has focused on how curricula can be designed to improve learning of a model on particular tasks. We consider the inverse problem: what can a curriculum tell us about how a learning system acquired a task? Using recurrent neural networks (RNNs) and models of common experimental neuroscience tasks, we demonstrate that curricula can be used to differentiate learning principles using target-based and a representation-based loss functions as use cases. In particular, we compare the performance of RNNs using target-based learning rules versus those using representational learning rules on three different curricula in the context of two tasks. We show that the learned state-space trajectories of RNNs trained by these two learning rules under all curricula tested are indistinguishable. However, by comparing learning times during different curricula, we can disambiguate the learning rules and challenge traditional approaches of interrogating learning systems. Although all animals in neuroscience lab settings are trained by curriculum-based procedures called shaping, almost no behavioral or neural data are collected or published on the relative successes or training times under different curricula. Our results motivate the systematic collection and curation of data during shaping by demonstrating curriculum learning in RNNs as a tool to probe and differentiate learning principles used by biological systems, over conventional statistical analyses of learned state spaces. | 1accept
|
Title: Learning to Infer. Abstract: Inference models, which replace an optimization-based inference procedure with a learned model, have been fundamental in advancing Bayesian deep learning, the most notable example being variational auto-encoders (VAEs). In this paper, we propose iterative inference models, which learn how to optimize a variational lower bound through repeatedly encoding gradients. Our approach generalizes VAEs under certain conditions, and by viewing VAEs in the context of iterative inference, we provide further insight into several recent empirical findings. We demonstrate the inference optimization capabilities of iterative inference models, explore unique aspects of these models, and show that they outperform standard inference models on typical benchmark data sets. | 0reject
|
Title: On the Adversarial Robustness of Vision Transformers. Abstract: Following the success in advancing natural language processing and understanding, transformers are expected to bring revolutionary changes to computer vision. This work provides the first and comprehensive study on the robustness of vision transformers (ViTs) against adversarial perturbations. Tested on various white-box and transfer attack settings, we find that ViTs possess better adversarial robustness when compared with convolutional neural networks (CNNs). This observation also holds for certified robustness. We summarize the following main observations contributing to the improved robustness of ViTs:
1) Features learned by ViTs contain less low-level information and are more generalizable, which contributes to superior robustness against adversarial perturbations.
2) Introducing convolutional or tokens-to-token blocks for learning low-level features in ViTs can improve classification accuracy but at the cost of adversarial robustness.
3) Increasing the proportion of transformers in the model structure (when the model consists of both transformer and CNN blocks) leads to better robustness. But for a pure transformer model, simply increasing the size or adding layers cannot guarantee a similar effect.
4) Pre-training on larger datasets does not significantly improve adversarial robustness though it is critical for training ViTs.
5) Adversarial training is also applicable to ViT for training robust models.
Furthermore, feature visualization and frequency analysis are conducted for explanation. The results show that ViTs are less sensitive to high-frequency perturbations than CNNs and there is a high correlation between how well the model learns low-level features and its robustness against different frequency-based perturbations. | 0reject
|
Title: Cross-Domain Cross-Set Few-Shot Learning via Learning Compact and Aligned Representations. Abstract: Few-shot learning (FSL) aims to recognize novel query examples with a small support set through leveraging prior knowledge learned from a large-scale training set. In this paper, we extend this task to a more practical setting where the domain shift exists between the support set and query examples and additional unlabeled data in the target domain can be adopted in the meta-training stage. Such new setting, termed cross-domain cross-set FSL (CDSC-FSL), requires the learning system not only to adapt to new classes with few examples but also to be consistent between different domains. To address this paradigm, we propose a novel approach, namely \textit{stab}PA, to learn prototypical compact and cross-domain aligned representations, so that domain shift and few-shot adaptation can be addressed simultaneously. We evaluate our approach on two new CDCS-FSL benchmarks adapted from the DomainNet and Office-Home datasets, respectively. Remarkably, our approach outperforms multiple elaborated baselines by a large margin and improves 5-shot accuracy by up to 4.7 points. | 2withdrawn
|
Title: Revisiting Knowledge Base Embedding as Tensor Decomposition. Abstract: We study the problem of knowledge base (KB) embedding, which is usually addressed through two frameworks---neural KB embedding and tensor decomposition. In this work, we theoretically analyze the neural embedding framework and subsequently connect it with tensor based embedding. Specifically, we show that in neural KB embedding the two commonly adopted optimization solutions---margin-based and negative sampling losses---are closely related to each other. We also reach the closed-form tensor that is implicitly approximated by popular neural KB approaches, revealing the underlying connection between neural and tensor based KB embedding models. Grounded in the theoretical results, we further present a tensor decomposition based framework KBTD to directly approximate the derived closed form tensor. Under this framework, the neural KB embedding models, such as NTN, TransE, Bilinear, and DISTMULT, are unified into a general tensor optimization architecture. Finally, we conduct experiments on the link prediction task in WordNet and Freebase, empirically demonstrating the effectiveness of the KBTD framework.
| 0reject
|
Title: BayesOpt Adversarial Attack. Abstract: Black-box adversarial attacks require a large number of attempts before finding successful adversarial examples that are visually indistinguishable from the original input. Current approaches relying on substitute model training, gradient estimation or genetic algorithms often require an excessive number of queries. Therefore, they are not suitable for real-world systems where the maximum query number is limited due to cost. We propose a query-efficient black-box attack which uses Bayesian optimisation in combination with Bayesian model selection to optimise over the adversarial perturbation and the optimal degree of search space dimension reduction. We demonstrate empirically that our method can achieve comparable success rates with 2-5 times fewer queries compared to previous state-of-the-art black-box attacks. | 1accept
|
Title: Stable cognitive maps for Path Integration emerge from fusing visual and proprioceptive sensors. Abstract: Spatial navigation in biological agents relies on the interplay between external (visual, olfactory, auditory, $\dots$) and proprioceptive (motor commands, linear and angular velocity, $\dots$) signals. How to combine and exploit these two streams of information, which vastly differ in terms of availability and reliability is a crucial issue. In the context of a new two--dimensional continuous environment we developed, we propose a direct-inverse model of environment dynamics to fuse image and action related signals, allowing reconstruction of the action relating the two successive images, as well as prediction of the new image from its current value and the action. The definition of those models naturally leads to the proposal of a minimalistic recurrent architecture, called Resetting Path Integrator (RPI), that can easily and reliably be trained to keep track of its position relative to its starting point during a sequence of movements. RPI updates its internal state using the (possibly noisy) proprioceptive signal, and occasionally resets it when the image signal is present. Notably, the internal state of this minimal model exhibits strong correlation with position in the environment due to the direct-inverse models, is stable across long trajectories through resetting, and allows for disambiguation of visually confusing positions in the environment through integration of past movement, making it a prime candidate for a \textbf{cognitive map}. Our architecture is compared to state-of-the-art LSTM networks on identical tasks, and consistently shows better performance while also offering more interpretable internal dynamics and higher-quality representations. | 0reject
|
Title: Count-Based Exploration with the Successor Representation. Abstract: The problem of exploration in reinforcement learning is well-understood in the tabular case and many sample-efficient algorithms are known. Nevertheless, it is often unclear how the algorithms in the tabular setting can be extended to tasks with large state-spaces where generalization is required. Recent promising developments generally depend on problem-specific density models or handcrafted features. In this paper we introduce a simple approach for exploration that allows us to develop theoretically justified algorithms in the tabular case but that also give us intuitions for new algorithms applicable to settings where function approximation is required. Our approach and its underlying theory is based on the substochastic successor representation, a concept we develop here. While the traditional successor representation is a representation that defines state generalization by the similarity of successor states, the substochastic successor representation is also able to implicitly count the number of times each state (or feature) has been observed. This extension connects two until now disjoint areas of research. We show in traditional tabular domains (RiverSwim and SixArms) that our algorithm empirically performs as well as other sample-efficient algorithms. We then describe a deep reinforcement learning algorithm inspired by these ideas and show that it matches the performance of recent pseudo-count-based methods in hard exploration Atari 2600 games. | 0reject
|
Title: Compressive Recovery Defense: A Defense Framework for $\ell_0, \ell_2$ and $\ell_\infty$ norm attacks.. Abstract: We provide recovery guarantees for compressible signals that have been corrupted with noise and extend the framework introduced in \cite{bafna2018thwarting} to defend neural networks against $\ell_0$, $\ell_2$, and $\ell_{\infty}$-norm attacks. In the case of $\ell_0$-norm noise, we provide recovery guarantees for Iterative Hard Thresholding (IHT) and Basis Pursuit (BP). For $\ell_2$-norm bounded noise, we provide recovery guarantees for BP, and for the case of $\ell_\infty$-norm bounded noise, we provide recovery guarantees for Dantzig Selector (DS). These guarantees theoretically bolster the defense framework introduced in \cite{bafna2018thwarting} for defending neural networks against adversarial inputs. Finally, we experimentally demonstrate the effectiveness of this defense framework against an array of $\ell_0$, $\ell_2$ and $\ell_\infty$-norm attacks. | 0reject
|
Title: Accurate and fast detection of copy number variations from short-read whole-genome sequencing with deep convolutional neural network. Abstract: A copy number variant (CNV) is a type of genetic mutation where a stretch of DNA is lost or duplicated once or multiple times. CNVs play important roles in the development of diseases and complex traits. CNV detection with short-read DNA sequencing technology is challenging because CNVs significantly vary in size and are similar to DNA sequencing artifacts. Many methods have been developed but still yield unsatisfactory results with high computational costs. Here, we propose CNV-Net, a novel approach for CNV detection using a six-layer convolutional neural network. We encode DNA sequencing information into RGB images and train the convolutional neural network with these images. The fitted convolutional neural network can then be used to predict CNVs from DNA sequencing data. We benchmark CNV-Net with two high-quality whole-genome sequencing datasets available from the Genome in a Bottle Consortium, considered as gold standard benchmarking datasets for CNV detection. We demonstrate that CNV-Net is more accurate and efficient in CNV detection than current tools. | 0reject
|
Title: Measuring the Intrinsic Dimension of Objective Landscapes. Abstract: Many recently trained neural networks employ large numbers of parameters to achieve good performance. One may intuitively use the number of parameters required as a rough gauge of the difficulty of a problem. But how accurate are such notions? How many parameters are really needed? In this paper we attempt to answer this question by training networks not in their native parameter space, but instead in a smaller, randomly oriented subspace. We slowly increase the dimension of this subspace, note at which dimension solutions first appear, and define this to be the intrinsic dimension of the objective landscape. The approach is simple to implement, computationally tractable, and produces several suggestive conclusions. Many problems have smaller intrinsic dimensions than one might suspect, and the intrinsic dimension for a given dataset varies little across a family of models with vastly different sizes. This latter result has the profound implication that once a parameter space is large enough to solve a problem, extra parameters serve directly to increase the dimensionality of the solution manifold. Intrinsic dimension allows some quantitative comparison of problem difficulty across supervised, reinforcement, and other types of learning where we conclude, for example, that solving the inverted pendulum problem is 100 times easier than classifying digits from MNIST, and playing Atari Pong from pixels is about as hard as classifying CIFAR-10. In addition to providing new cartography of the objective landscapes wandered by parameterized models, the method is a simple technique for constructively obtaining an upper bound on the minimum description length of a solution. A byproduct of this construction is a simple approach for compressing networks, in some cases by more than 100 times. | 1accept
|
Title: Gradient Descent Ascent for Min-Max Problems on Riemannian Manifolds. Abstract: In the paper, we study a class of useful non-convex minimax optimization problems on Riemanian manifolds and propose a class of Riemanian gradient descent ascent algorithms to solve these minimax problems. Specifically, we propose a new
Riemannian gradient descent ascent (RGDA) algorithm for the deterministic minimax optimization.
Moreover, we prove that the RGDA has a sample complexity of $O(\kappa^2\epsilon^{-2})$ for finding an $\epsilon$-stationary point of the nonconvex strongly-concave minimax problems, where $\kappa$ denotes the condition number.
At the same time, we introduce a Riemannian stochastic gradient descent ascent (RSGDA) algorithm for the stochastic minimax optimization. In the theoretical analysis, we prove that the RSGDA can achieve a sample complexity of $O(\kappa^4\epsilon^{-4})$.
To further reduce the sample complexity, we propose a novel momentum variance-reduced Riemannian stochastic gradient descent ascent (MVR-RSGDA) algorithm based on a new momentum variance-reduced technique of STORM. We prove that the MVR-RSGDA algorithm achieves a lower sample complexity of $\tilde{O}(\kappa^{4}\epsilon^{-3})$ without large batches, which reaches near the best known sample complexity for its Euclidean counterparts. Extensive experimental results on the robust deep neural networks training over Stiefel manifold demonstrate the efficiency of our proposed algorithms. | 0reject
|
Title: Characterizing convolutional neural networks with one-pixel signature. Abstract: We propose a new representation, one-pixel signature, that can be used to reveal the characteristics of the convolution neural networks (CNNs). Here, each CNN classifier is associated with a signature that is created by generating, pixel-by-pixel, an adversarial value that is the result of the largest change to the class prediction. The one-pixel signature is agnostic to the design choices of CNN architectures such as type, depth, activation function, and how they were trained. It can be computed efficiently for a black-box classifier without accessing the network parameters. Classic networks such as LetNet, VGG, AlexNet, and ResNet demonstrate different characteristics in their signature images. For application, we focus on the classifier backdoor detection problem where a CNN classifier has been maliciously inserted with an unknown Trojan. We show the effectiveness of the one-pixel signature in detecting backdoored CNN. Our proposed one-pixel signature representation is general and it can be applied in problems where discriminative classifiers, particularly neural network based, are to be characterized. | 2withdrawn
|
Title: Kanerva++: Extending the Kanerva Machine With Differentiable, Locally Block Allocated Latent Memory. Abstract: Episodic and semantic memory are critical components of the human memory model. The theory of complementary learning systems (McClelland et al., 1995) suggests that the compressed representation produced by a serial event (episodic memory) is later restructured to build a more generalized form of reusable knowledge (semantic memory). In this work, we develop a new principled Bayesian memory allocation scheme that bridges the gap between episodic and semantic memory via a hierarchical latent variable model. We take inspiration from traditional heap allocation and extend the idea of locally contiguous memory to the Kanerva Machine, enabling a novel differentiable block allocated latent memory. In contrast to the Kanerva Machine, we simplify the process of memory writing by treating it as a fully feed forward deterministic process, relying on the stochasticity of the read key distribution to disperse information within the memory. We demonstrate that this allocation scheme improves performance in memory conditional image generation, resulting in new state-of-the-art conditional likelihood values on binarized MNIST (≤41.58 nats/image) , binarized Omniglot (≤66.24 nats/image), as well as presenting competitive performance on CIFAR10, DMLab Mazes, Celeb-A and ImageNet32×32. | 1accept
|
Title: Towards Impartial Multi-task Learning. Abstract: Multi-task learning (MTL) has been widely used in representation learning. However, naively training all tasks simultaneously may lead to the partial training issue, where specific tasks are trained more adequately than others. In this paper, we propose to learn multiple tasks impartially. Specifically, for the task-shared parameters, we optimize the scaling factors via a closed-form solution, such that the aggregated gradient (sum of raw gradients weighted by the scaling factors) has equal projections onto individual tasks. For the task-specific parameters, we dynamically weigh the task losses so that all of them are kept at a comparable scale. Further, we find the above gradient balance and loss balance are complementary and thus propose a hybrid balance method to further improve the performance. Our impartial multi-task learning (IMTL) can be end-to-end trained without any heuristic hyper-parameter tuning, and is general to be applied on all kinds of losses without any distribution assumption. Moreover, our IMTL can converge to similar results even when the task losses are designed to have different scales, and thus it is scale-invariant. We extensively evaluate our IMTL on the standard MTL benchmarks including Cityscapes, NYUv2 and CelebA. It outperforms existing loss weighting methods under the same experimental settings. | 1accept
|
Title: Deep Randomized Least Squares Value Iteration. Abstract: Exploration while learning representations is one of the main challenges Deep
Reinforcement Learning (DRL) faces today. As the learned representation is dependant in the observed data, the exploration strategy has a crucial role. The popular DQN algorithm has improved significantly the capabilities of Reinforcement
Learning (RL) algorithms to learn state representations from raw data, yet, it uses
a naive exploration strategy which is statistically inefficient. The Randomized
Least Squares Value Iteration (RLSVI) algorithm (Osband et al., 2016), on the
other hand, explores and generalizes efficiently via linearly parameterized value
functions. However, it is based on hand-designed state representation that requires
prior engineering work for every environment. In this paper, we propose a Deep
Learning adaptation for RLSVI. Rather than using hand-design state representation, we use a state representation that is being learned directly from the data by a
DQN agent. As the representation is being optimized during the learning process,
a key component for the suggested method is a likelihood matching mechanism,
which adapts to the changing representations. We demonstrate the importance of
the various properties of our algorithm on a toy problem and show that our method
outperforms DQN in five Atari benchmarks, reaching competitive results with the
Rainbow algorithm. | 0reject
|
Title: Soft Q-Learning with Mutual-Information Regularization. Abstract: We propose a reinforcement learning (RL) algorithm that uses mutual-information regularization to optimize a prior action distribution for better performance and exploration. Entropy-based regularization has previously been shown to improve both exploration and robustness in challenging sequential decision-making tasks. It does so by encouraging policies to put probability mass on all actions. However, entropy regularization might be undesirable when actions have significantly different importance. In this paper, we propose a theoretically motivated framework that dynamically weights the importance of actions by using the mutual-information. In particular, we express the RL problem as an inference problem where the prior probability distribution over actions is subject to optimization. We show that the prior optimization introduces a mutual-information regularizer in the RL objective. This regularizer encourages the policy to be close to a non-uniform distribution that assigns higher probability mass to more important actions. We empirically demonstrate that our method significantly improves over entropy regularization methods and unregularized methods. | 1accept
|
Title: Graph Classification with Geometric Scattering. Abstract: One of the most notable contributions of deep learning is the application of convolutional neural networks (ConvNets) to structured signal classification, and in particular image classification. Beyond their impressive performances in supervised learning, the structure of such networks inspired the development of deep filter banks referred to as scattering transforms. These transforms apply a cascade of wavelet transforms and complex modulus operators to extract features that are invariant to group operations and stable to deformations. Furthermore, ConvNets inspired recent advances in geometric deep learning, which aim to generalize these networks to graph data by applying notions from graph signal processing to learn deep graph filter cascades. We further advance these lines of research by proposing a geometric scattering transform using graph wavelets defined in terms of random walks on the graph. We demonstrate the utility of features extracted with this designed deep filter bank in graph classification of biochemistry and social network data (incl. state of the art results in the latter case), and in data exploration, where they enable inference of EC exchange preferences in enzyme evolution. | 0reject
|
Title: Latent Convergent Cross Mapping. Abstract: Discovering causal structures of temporal processes is a major tool of scientific inquiry because it helps us better understand and explain the mechanisms driving a phenomenon of interest, thereby facilitating analysis, reasoning, and synthesis for such systems.
However, accurately inferring causal structures within a phenomenon based on observational data only is still an open problem. Indeed, this type of data usually consists in short time series with missing or noisy values for which causal inference is increasingly difficult. In this work, we propose a method to uncover causal relations in chaotic dynamical systems from short, noisy and sporadic time series (that is, incomplete observations at infrequent and irregular intervals) where the classical convergent cross mapping (CCM) fails. Our method works by learning a Neural ODE latent process modeling the state-space dynamics of the time series and by checking the existence of a continuous map between the resulting processes. We provide theoretical analysis and show empirically that Latent-CCM can reliably uncover the true causal pattern, unlike traditional methods. | 1accept
|
Title: Tractable Dendritic RNNs for Identifying Unknown Nonlinear Dynamical Systems. Abstract: In many scientific disciplines, we are interested in inferring the nonlinear dynamical system underlying a set of observed time series, a challenging task in the face of chaotic behavior and noise. Previous deep learning approaches toward this goal often suffered from a lack of interpretability and tractability. In particular, the high-dimensional latent spaces often required for a faithful embedding, even when the underlying dynamics lives on a lower-dimensional manifold, can hamper theoretical analysis. Motivated by the emerging principles of dendritic computation, we augment a dynamically interpretable and mathematically tractable piecewise-linear (PL) recurrent neural network (RNN) by a linear spline basis expansion. We show that this approach retains all the theoretically appealing properties of the simple PLRNN, yet boosts its capacity for approximating arbitrary nonlinear dynamical systems in comparatively low dimensions. We introduce two frameworks for training the system, one based on fast and scalable variational inference, and another combining BPTT with teacher forcing. We show that the dendritically expanded PLRNN achieves better reconstructions with fewer parameters and dimensions on various dynamical systems benchmarks and compares favorably to other methods, while retaining a tractable and interpretable structure. | 0reject
|
Title: Autoencoders and Generative Adversarial Networks for Imbalanced Sequence Classification. Abstract: We introduce a novel synthetic oversampling method for variable length, multi- feature sequence datasets based on autoencoders and generative adversarial net- works. We show that this method improves classification accuracy for highly imbalanced sequence classification tasks. We show that this method outperforms standard oversampling techniques that use techniques such as SMOTE and autoencoders. We also use generative adversarial networks on the majority class as an outlier detection method for novelty detection, with limited classification improvement. We show that the use of generative adversarial network based synthetic data improves classification model performance on a variety of sequence data sets.
| 0reject
|
Title: Value-aware transformers for 1.5d data. Abstract: Sparse sequential highly-multivariate data of the form characteristic of hospital in-patient investigation and treatment poses a considerable challenge for representation learning. Such data is neither faithfully reducible to 1d nor dense enough to constitute multivariate series. Conventional models compromise their data by requiring these forms at the point of input. Building on contemporary sequence-modelling architectures we design a value-aware transformer, prompting a reconceptualisation of our data as 1.5-dimensional: a token-value form both respecting its sequential nature and augmenting it with a quantifier. Experiments focused on sequential in-patient laboratory data up to 48hrs after hospital admission show that the value-aware transformer performs favourably versus competitive baselines on in-hospital mortality and length-of-stay prediction within the MIMIC-III dataset. | 0reject
|
Title: LatTe Flows: Latent Temporal Flows for Multivariate Sequence Analysis . Abstract: We introduce Latent Temporal Flows (\emph{LatTe-Flows}), a method for probabilistic multivariate time-series analysis tailored for high dimensional systems whose temporal dynamics are driven by variations in a lower-dimensional discriminative subspace. We perform indirect learning from hidden traits of observed sequences by assuming that the random vector representing the data is generated from an unobserved low-dimensional latent vector. \emph{LatTe-Flows} jointly learns auto-encoder mappings to a latent space and learns the temporal distribution of lower-dimensional embeddings of input sequences. Since encoder networks retain only the essential information to generate a latent manifold, the temporal distribution transitions can be more efficiently uncovered by time conditioned Normalizing Flows. The learned latent effects can then be directly transferred into the observed space through the decoder network. We demonstrate that the proposed method significantly outperforms the state-of-the-art on multi-step forecasting benchmarks, while enjoying reduced computational complexity on several real-world datasets. We apply {\emph{LatTe-Flows}} to a challenging sensor-signal forecasting task, using multivariate time-series measurements collected by wearable devices, an increasingly relevant health application.
| 0reject
|
Title: Black Box Recursive Translations for Molecular Optimization. Abstract: Machine learning algorithms for generating molecular structures offer a promising new approach to drug discovery. We cast molecular optimization as a translation problem, where the goal is to map an input compound to a target compound with improved biochemical properties. Remarkably, we observe that when generated molecules are iteratively fed back into the translator, molecular compound attributes improve with each step. We show that this finding is invariant to the choice of translation model, making this a "black box" algorithm. We call this method Black Box Recursive Translation (BBRT), a new inference method for molecular property optimization. This simple, powerful technique operates strictly on the inputs and outputs of any translation model. We obtain new state-of-the-art results for molecular property optimization tasks using our simple drop-in replacement with well-known sequence and graph-based models. Our method provides a significant boost in performance relative to its non-recursive peers with just a simple "``for" loop. Further, BBRT is highly interpretable, allowing users to map the evolution of newly discovered compounds from known starting points. | 0reject
|
Title: MobileBERT: Task-Agnostic Compression of BERT by Progressive Knowledge Transfer. Abstract: The recent development of Natural Language Processing (NLP) has achieved great success using large pre-trained models with hundreds of millions of parameters. However, these models suffer from the heavy model size and high latency such that we cannot directly deploy them to resource-limited mobile devices. In this paper, we propose MobileBERT for compressing and accelerating the popular BERT model. Like BERT, MobileBERT is task-agnostic; that is, it can be universally applied to various downstream NLP tasks via fine-tuning. MobileBERT is a slimmed version of BERT-LARGE augmented with bottleneck structures and a carefully designed balance between self-attentions and feed-forward networks. To train MobileBERT, we use a bottom-to-top progressive scheme to transfer the intrinsic knowledge of a specially designed Inverted Bottleneck BERT-LARGE teacher to it. Empirical studies show that MobileBERT is 4.3x smaller and 4.0x faster than original BERT-BASE while achieving competitive results on well-known NLP benchmarks. On the natural language inference tasks of GLUE, MobileBERT achieves 0.6 GLUE score performance degradation, and 367 ms latency on a Pixel 3 phone. On the SQuAD v1.1/v2.0 question answering task, MobileBERT achieves a 90.0/79.2 dev F1 score, which is 1.5/2.1 higher than BERT-BASE. | 2withdrawn
|
Title: Semantics Preserving Adversarial Attacks. Abstract: While progress has been made in crafting visually imperceptible adversarial examples, constructing semantically meaningful ones remains a challenge. In this paper, we propose a framework to generate semantics preserving adversarial examples. First, we present a manifold learning method to capture the semantics of the inputs. The motivating principle is to learn the low-dimensional geometric summaries of the inputs via statistical inference. Then, we perturb the elements of the learned manifold using the Gram-Schmidt process to induce the perturbed elements to remain in the manifold. To produce adversarial examples, we propose an efficient algorithm whereby we leverage the semantics of the inputs as a source of knowledge upon which we impose adversarial constraints. We apply our approach on toy data, images and text, and show its effectiveness in producing semantics preserving adversarial examples which evade existing defenses against adversarial attacks. | 0reject
|
Title: Co-Attentive Equivariant Neural Networks: Focusing Equivariance On Transformations Co-Occurring in Data. Abstract: Equivariance is a nice property to have as it produces much more parameter efficient neural architectures and preserves the structure of the input through the feature mapping. Even though some combinations of transformations might never appear (e.g. an upright face with a horizontal nose), current equivariant architectures consider the set of all possible transformations in a transformation group when learning feature representations. Contrarily, the human visual system is able to attend to the set of relevant transformations occurring in the environment and utilizes this information to assist and improve object recognition. Based on this observation, we modify conventional equivariant feature mappings such that they are able to attend to the set of co-occurring transformations in data and generalize this notion to act on groups consisting of multiple symmetries. We show that our proposed co-attentive equivariant neural networks consistently outperform conventional rotation equivariant and rotation & reflection equivariant neural networks on rotated MNIST and CIFAR-10. | 1accept
|
Title: Distance-based Composable Representations with Neural Networks. Abstract: We introduce a new deep learning technique that builds individual and class representations based on distance estimates to randomly generated contextual dimensions for different modalities. Recent works have demonstrated advantages to creating representations from probability distributions over their contexts rather than single points in a low-dimensional Euclidean vector space. These methods, however, rely on pre-existing features and are limited to textual information. In this work, we obtain generic template representations that are vectors containing the average distance of a class to randomly generated contextual information. These representations have the benefit of being both interpretable and composable. They are initially learned by estimating the Wasserstein distance for different data subsets with deep neural networks. Individual samples or instances can then be compared to the generic class representations, which we call templates, to determine their similarity and thus class membership. We show that this technique, which we call WDVec, delivers good results for multi-label image classification. Additionally, we illustrate the benefit of templates and their composability by performing retrieval with complex queries where we modify the information content in the representations. Our method can be used in conjunction with any existing neural network and create theoretically infinitely large feature maps. | 0reject
|
Title: Reintroducing Straight-Through Estimators as Principled Methods for Stochastic Binary Networks. Abstract: Training neural networks with binary weights and activations is a challenging problem due to the lack of gradients and difficulty of optimization over discrete weights.
Many successful experimental results have been achieved with empirical straight-through (ST) approaches, proposing a variety of ad-hoc rules for propagating gradients through non-differentiable activations and updating discrete weights. At the same time, ST methods can be truly derived as estimators in the stochastic binary network (SBN) model with Bernoulli weights. We advance these derivations to a more complete and systematic study. We analyze properties, estimation accuracy, obtain different forms of correct ST estimators for activations and weights, explain existing empirical approaches and their shortcomings, explain how latent weights arise from the mirror descent method when optimizing over probabilities. This allows to reintroduce, once empirical, ST methods as sound approximations, apply them with clarity and develop further improvements. | 0reject
|
Title: Graph Traversal with Tensor Functionals: A Meta-Algorithm for Scalable Learning. Abstract: Graph Representation Learning (GRL) methods have impacted fields from chemistry to social science. However, their algorithmic implementations are specialized to specific use-cases e.g. "message passing" methods are run differently from "node embedding" ones. Despite their apparent differences, all these methods utilize the graph structure, and therefore, their learning can be approximated with stochastic graph traversals. We propose Graph Traversal via Tensor Functionals (GTTF), a unifying meta-algorithm framework for easing the implementation of diverse graph algorithms and enabling transparent and efficient scaling to large graphs. GTTF is founded upon a data structure (stored as a sparse tensor) and a stochastic graph traversal algorithm (described using tensor operations). The algorithm is a functional that accept two functions, and can be specialized to obtain a variety of GRL models and objectives, simply by changing those two functions. We show for a wide class of methods, our algorithm learns in an unbiased fashion and, in expectation, approximates the learning as if the specialized implementations were run directly.
With these capabilities, we scale otherwise non-scalable methods to set state-of-the-art on large graph datasets while being more efficient than existing GRL libraries -- with only a handful of lines of code for each method specialization. | 1accept
|
Title: Training Federated GANs with Theoretical Guarantees: A Universal Aggregation Approach. Abstract: Recently, Generative Adversarial Networks (GANs) have demonstrated their potential in federated learning, i.e., learning a centralized model from data privately hosted by multiple sites. A federated GAN jointly trains a centralized generator and multiple private discriminators hosted at different sites. A major theoretical challenge for the federated GAN is the heterogeneity of the local data distributions. Traditional approaches cannot guarantee to learn the target distribution, which is a mixture of the highly different local distributions. This paper tackles this theoretical challenge, and for the first time, provides a provably correct framework for federated GAN. We propose a new approach called Universal Aggregation, which simulates a centralized discriminator via carefully aggregating the mixture of all private discriminators. We prove that a generator trained with this simulated centralized discriminator can learn the desired target distribution. Through synthetic and real datasets, we show that our method can learn the mixture of largely different distributions, when existing federated GAN methods fail to. | 0reject
|
Title: Differentiable Optimization of Generalized Nondecomposable Functions using Linear Programs. Abstract: We propose a framework which makes it feasible to directly train deep neural networks with respect to popular families of task-specific non-decomposable per- formance measures such as AUC, multi-class AUC, F -measure and others, as well as models such as non-negative matrix factorization. A common feature of the optimization model that emerges from these tasks is that it involves solving a Linear Programs (LP) during training where representations learned by upstream layers influence the constraints. The constraint matrix is not only large but the constraints are also modified at each iteration. We show how adopting a set of influential ideas proposed by Mangasarian for 1-norm SVMs – which advocates for solving LPs with a generalized Newton method – provides a simple and effective solution. In particular, this strategy needs little unrolling, which makes it more efficient during backward pass. While a number of specialized algorithms have been proposed for the models that we de- scribe here, our module turns out to be applicable without any specific adjustments or relaxations. We describe each use case, study its properties and demonstrate the efficacy of the approach over alternatives which use surrogate lower bounds and often, specialized optimization schemes. Frequently, we achieve superior computational behavior and performance improvements on common datasets used in the literature.
| 0reject
|
Title: The Deep Weight Prior. Abstract: Bayesian inference is known to provide a general framework for incorporating prior knowledge or specific properties into machine learning models via carefully choosing a prior distribution. In this work, we propose a new type of prior distributions for convolutional neural networks, deep weight prior (DWP), that exploit generative models to encourage a specific structure of trained convolutional filters e.g., spatial correlations of weights. We define DWP in the form of an implicit distribution and propose a method for variational inference with such type of implicit priors. In experiments, we show that DWP improves the performance of Bayesian neural networks when training data are limited, and initialization of weights with samples from DWP accelerates training of conventional convolutional neural networks.
| 1accept
|
Title: Interpreting Molecule Generative Models for Interactive Molecule Discovery. Abstract: Discovering novel molecules with desired properties is crucial for advancing drug discovery and chemical science. Recently deep generative models can synthesize new molecules by sampling random vectors from latent space and then decoding them to a molecule structure. However, through the feedforward generation pipeline, it is difficult to reveal the underlying connections between latent space and molecular properties as well as customize the output molecule with desired properties. In this work, we develop a simple yet effective method to interpret the latent space of the learned generative models with various molecular properties for more interactive molecule generation and discovery. This method, called Molecular Space Explorer (MolSpacE), is model-agnostic and can work with any pre-trained molecule generative models in an off-the-shelf manner. It first identifies latent directions that govern certain molecular properties via the property separation hyperplane and then moves molecules along the directions for smooth change of molecular structures and properties. This method achieves interactive molecule discovery through identifying interpretable and steerable concepts that emerge in the representations of generative models. Experiments show that MolSpacE can manipulate the output molecule toward desired properties with high success. We further quantify and compare the interpretability of multiple state-of-the-art molecule generative models. An interface and a demo video are developed to illustrate the promising application of interactive molecule discovery. | 0reject
|
Title: Learning Document Embeddings With CNNs. Abstract: This paper proposes a new model for document embedding. Existing approaches either require complex inference or use recurrent neural networks that are difficult to parallelize. We take a different route and use recent advances in language modeling to develop a convolutional neural network embedding model. This allows us to train deeper architectures that are fully parallelizable. Stacking layers together increases the receptive filed allowing each successive layer to model increasingly longer range semantic dependences within the document. Empirically we demonstrate superior results on two publicly available benchmarks. Full code will be released with the final version of this paper. | 0reject
|
Title: Hierarchical RL Using an Ensemble of Proprioceptive Periodic Policies. Abstract: In this paper we introduce a simple, robust approach to hierarchically training an agent in the setting of sparse reward tasks.
The agent is split into a low-level and a high-level policy. The low-level policy only accesses internal, proprioceptive dimensions of the state observation. The low-level policies are trained with a simple reward that encourages changing the values of the non-proprioceptive dimensions. Furthermore, it is induced to be periodic with the use a ``phase function.'' The high-level policy is trained using a sparse, task-dependent reward, and operates by choosing which of the low-level policies to run at any given time. Using this approach, we solve difficult maze and navigation tasks with sparse rewards using the Mujoco Ant and Humanoid agents and show improvement over recent hierarchical methods. | 1accept
|
Title: Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networks. Abstract: We consider the problem of detecting out-of-distribution images in neural networks. We propose ODIN, a simple and effective method that does not require any change to a pre-trained neural network. Our method is based on the observation that using temperature scaling and adding small perturbations to the input can separate the softmax score distributions of in- and out-of-distribution images, allowing for more effective detection. We show in a series of experiments that ODIN is compatible with diverse network architectures and datasets. It consistently outperforms the baseline approach by a large margin, establishing a new state-of-the-art performance on this task. For example, ODIN reduces the false positive rate from the baseline 34.7% to 4.3% on the DenseNet (applied to CIFAR-10 and Tiny-ImageNet) when the true positive rate is 95%. | 1accept
|
Title: Fuzzy Tiling Activations: A Simple Approach to Learning Sparse Representations Online. Abstract: Recent work has shown that sparse representations---where only a small percentage of units are active---can significantly reduce interference. Those works, however, relied on relatively complex regularization or meta-learning approaches, that have only been used offline in a pre-training phase. In this work, we pursue a direction that achieves sparsity by design, rather than by learning. Specifically, we design an activation function that produces sparse representations deterministically by construction, and so is more amenable to online training. The idea relies on the simple approach of binning, but overcomes the two key limitations of binning: zero gradients for the flat regions almost everywhere, and lost precision---reduced discrimination---due to coarse aggregation. We introduce a Fuzzy Tiling Activation (FTA) that provides non-negligible gradients and produces overlap between bins that improves discrimination. We first show that FTA is robust under covariate shift in a synthetic online supervised learning problem, where we can vary the level of correlation and drift. Then we move to the deep reinforcement learning setting and investigate both value-based and policy gradient algorithms that use neural networks with FTAs, in classic discrete control and Mujoco continuous control environments. We show that algorithms equipped with FTAs are able to learn a stable policy faster without needing target networks on most domains. | 1accept
|
Title: A Simple Dynamic Learning Rate Tuning Algorithm For Automated Training of DNNs. Abstract: Training neural networks on image datasets generally require extensive experimentation to find the optimal learning rate regime. Especially, for the cases of adversarial training or for training a newly synthesized model, one would not know the best learning rate regime beforehand. We propose an automated algorithm for determining the learning rate trajectory, that works across datasets and models for both natural and adversarial training, without requiring any dataset/model specific tuning. It is a stand-alone, parameterless, adaptive approach with no computational overhead. We theoretically discuss the algorithm's convergence behavior. We empirically validate our algorithm extensively. Our results show that our proposed approach \emph{consistently} achieves top-level accuracy compared to SOTA baselines in the literature in natural training, as well as in adversarial training. | 0reject
|
Title: FERMI: Fair Empirical Risk Minimization via Exponential Rényi Mutual Information. Abstract: Several notions of fairness, such as demographic parity and equal opportunity, are defined based on statistical independence between a predicted target and a sensitive attribute. In machine learning applications, however, the data distribution is unknown to the learner and statistical independence is not verifiable. Hence, the learner could only resort to empirical evaluation of the degree of fairness violation. Many fairness violation notions are defined as a divergence/distance between the joint distribution of the target and sensitive attributes and the Kronecker product of their marginals, such as \Renyi correlation, mutual information, $L_\infty$ distance, to name a few.
In this paper, we propose another notion of fairness violation, called Exponential R\'enyi Mutual Information (ERMI) between sensitive attributes and the predicted target. We show that ERMI is a strong fairness violation notion in the sense that it provides an upper bound guarantee on all of the aforementioned notions of fairness violation. We also propose the Fair Empirical Risk Minimization via ERMI regularization framework, called FERMI. Whereas existing in-processing fairness algorithms are deterministic, we provide a stochastic optimization method for solving FERMI that is amenable to large-scale problems. In addition, we provide a batch (deterministic) method to solve FERMI. Both of our proposed algorithms come with theoretical convergence guarantees. Our experiments show that FERMI achieves the most favorable tradeoffs between fairness violation and accuracy on test data across different problem setups, even when fairness violation is measured in notions other than ERMI. | 0reject
|
Title: Empirical confidence estimates for classification by deep neural networks. Abstract: How well can we estimate the probability that the classification predicted by a deep neural network is correct (or in the Top 5)? It is well-known that the softmax values of the network are not estimates of the probabilities of class labels. However, there is a misconception that these values are not informative. We define the notion of implied loss and prove that if an uncertainty measure is an implied loss, then low uncertainty means high probability of correct (or Top-k) classification on the test set. We demonstrate empirically that these values can be used to measure the confidence that the classification is correct. Our method is simple to use on existing networks: we proposed confidence measures for Top-k which can be evaluated by binning values on the test set. | 0reject
|
Title: CKConv: Continuous Kernel Convolution For Sequential Data. Abstract: Conventional neural architectures for sequential data present important limitations. Recurrent neural networks suffer from exploding and vanishing gradients, small effective memory horizons, and must be trained sequentially. Convolutional neural networks cannot handle sequences of unknown size and their memory horizon must be defined a priori. In this work, we show that these problems can be solved by formulating the convolutional kernels of CNNs as continuous functions. The resulting Continuous Kernel Convolution (CKConv) handles arbitrarily long sequences in a parallel manner, within a single operation, and without relying on any form of recurrence. We show that Continuous Kernel Convolutional Networks (CKCNNs) obtain state-of-the-art results in multiple datasets, e.g., permuted MNIST, and, thanks to their continuous nature, are able to handle non-uniformly sampled datasets and irregularly-sampled data natively. CKCNNs match or perform better than neural ODEs designed for these purposes in a faster and simpler manner. | 1accept
|
Title: Sample Efficient Deep Neuroevolution in Low Dimensional Latent Space. Abstract: Current deep neuroevolution models are usually trained in a large parameter search space for complex learning tasks, e.g. playing video games, which needs billions of samples and thousands of search steps to obtain significant performance. This raises a question of whether we can make use of sequential data generated during evolution, encode input samples, and evolve in low dimensional parameter space with latent state input in a fast and efficient manner. Here we give an affirmative answer: we train a VAE to encode input samples, then an RNN to model environment dynamics and handle temporal information, and last evolve our low dimensional policy network in latent space. We demonstrate that this approach is surprisingly efficient: our experiments on Atari games show that within 10M frames and 30 evolution steps of training, our algorithm could achieve competitive result compared with ES, A3C, and DQN which need billions of frames. | 0reject
|
Title: Implicit vs Unfolded Graph Neural Networks. Abstract: It has been observed that graph neural networks (GNN) sometimes struggle to maintain a healthy balance between modeling long-range dependencies across nodes while avoiding unintended consequences such as oversmoothed node representations. To address this issue (among other things), two separate strategies have recently been proposed, namely implicit and unfolded GNNs. The former treats node representations as the fixed points of a deep equilibrium model that can efficiently facilitate arbitrary implicit propagation across the graph with a fixed memory footprint. In contrast, the latter involves treating graph propagation as the unfolded descent iterations as applied to some graph-regularized energy function. While motivated differently, in this paper we carefully elucidate the similarity and differences of these methods, quantifying explicit situations where the solutions they produced may actually be equivalent and others where behavior diverges. This includes the analysis of convergence, representational capacity, and interpretability. We also provide empirical head-to-head comparisons across a variety of synthetic and public real-world benchmarks. | 2withdrawn
|
Title: ChainGAN: A sequential approach to GANs. Abstract: We propose a new architecture and training methodology for generative adversarial networks. Current approaches attempt to learn the transformation from a noise sample to a generated data sample in one shot. Our proposed generator architecture, called ChainGAN, uses a two-step process. It first attempts to transform a noise vector into a crude sample, similar to a traditional generator. Next, a chain of networks, called editors, attempt to sequentially enhance this sample. We train each of these units independently, instead of with end-to-end backpropagation on the entire chain. Our model is robust, efficient, and flexible as we can apply it to various network architectures. We provide rationale for our choices and experimentally evaluate our model, achieving competitive results on several datasets. | 0reject
|
Title: Log Hyperbolic Cosine Loss Improves Variational Auto-Encoder. Abstract: In Variational Auto-Encoder (VAE), the default choice of reconstruction loss function between the decoded sample and the input is the squared $L_2$. We propose to replace it with the log hyperbolic cosine (log-cosh) loss, which behaves as $L_2$ at small values and as $L_1$ at large values, and differentiable everywhere. Compared with $L_2$, the log-cosh loss improves the reconstruction without damaging the latent space optimization, thus automatically keeping a balance between the reconstruction and the generation. Extensive experiments on MNIST and CelebA datasets show that the log-cosh reconstruction loss significantly improves the performance of VAE and its variants in output quality, measured by sharpness and FID score. In addition, the gradient of the log-cosh is a simple tanh function, which makes the implementation of gradient descent as simple as adding one sentence in coding. | 0reject
|
Title: PIVQGAN: Posture and Identity Disentangled Image-to-Image Translation via Vector Quantization. Abstract: One popular objective for the image-to-image translation task is to independently control the coarse-level object arrangements (posture) and the fine-grained level styling (identity) of the generated image from two exemplar sources. To approach this objective, we propose PIVQGAN with two novel techniques in the framework of StyleGAN2. First, we propose a Vector-Quantized Spatial Normalization (VQSN) module for the generator for better pose-identity disentanglement. The VQSN module automatically learns to encode the shaping and composition information from the commonly shared objects inside the training-set images. Second, we design a joint-training scheme with self-supervision methods for the GAN-Inversion encoder and the generator. Specifically, we let the encoder and generator reconstruct images from two differently augmented variants of the original ones, one defining the pose and the other for identity. The VQSN module facilitates a more delicate separation of posture and identity, while the training scheme ensures the VQSN module learns the pose-related representations. Comprehensive experiments conducted on various datasets show better synthesis image quality and disentangling scores of our model. Moreover, we present model applications beyond posture-identity disentangling, thanks to the latent-space reducing feature of the leveraged VQSN module. | 0reject
|
Title: Learning the Arrow of Time for Problems in Reinforcement Learning. Abstract: We humans have an innate understanding of the asymmetric progression of time, which we use to efficiently and safely perceive and manipulate our environment. Drawing inspiration from that, we approach the problem of learning an arrow of time in a Markov (Decision) Process. We illustrate how a learned arrow of time can capture salient information about the environment, which in turn can be used to measure reachability, detect side-effects and to obtain an intrinsic reward signal. Finally, we propose a simple yet effective algorithm to parameterize the problem at hand and learn an arrow of time with a function approximator (here, a deep neural network). Our empirical results span a selection of discrete and continuous environments, and demonstrate for a class of stochastic processes that the learned arrow of time agrees reasonably well with a well known notion of an arrow of time due to Jordan, Kinderlehrer and Otto (1998). | 1accept
|
Title: Learning Corresponded Rationales for Text Matching. Abstract: The ability to predict matches between two sources of text has a number of applications including natural language inference (NLI) and question answering (QA). While flexible neural models have become effective tools in solving these tasks, they are rarely transparent in terms of the mechanism that mediates the prediction. In this paper, we propose a self-explaining architecture where the model is forced to highlight, in a dependent manner, how spans of one side of the input match corresponding segments of the other side in order to arrive at the overall decision. The text spans are regularized to be coherent and concise, and their correspondence is captured explicitly. The text spans -- rationales -- are learned entirely as latent mechanisms, guided only by the distal supervision from the end-to-end task. We evaluate our model on both NLI and QA using three publicly available datasets. Experimental results demonstrate quantitatively and qualitatively that our method delivers interpretable justification of the prediction without sacrificing state-of-the-art performance. Our code and data split will be publicly available. | 0reject
|
Title: Neural Mechanics: Symmetry and Broken Conservation Laws in Deep Learning Dynamics. Abstract: Understanding the dynamics of neural network parameters during training is one of the key challenges in building a theoretical foundation for deep learning. A central obstacle is that the motion of a network in high-dimensional parameter space undergoes discrete finite steps along complex stochastic gradients derived from real-world datasets. We circumvent this obstacle through a unifying theoretical framework based on intrinsic symmetries embedded in a network's architecture that are present for any dataset. We show that any such symmetry imposes stringent geometric constraints on gradients and Hessians, leading to an associated conservation law in the continuous-time limit of stochastic gradient descent (SGD), akin to Noether's theorem in physics. We further show that finite learning rates used in practice can actually break these symmetry induced conservation laws. We apply tools from finite difference methods to derive modified gradient flow, a differential equation that better approximates the numerical trajectory taken by SGD at finite learning rates. We combine modified gradient flow with our framework of symmetries to derive exact integral expressions for the dynamics of certain parameter combinations. We empirically validate our analytic expressions for learning dynamics on VGG-16 trained on Tiny ImageNet. Overall, by exploiting symmetry, our work demonstrates that we can analytically describe the learning dynamics of various parameter combinations at finite learning rates and batch sizes for state of the art architectures trained on any dataset. | 1accept
|
Title: Total Style Transfer with a Single Feed-Forward Network. Abstract: Recent image style transferring methods achieved arbitrary stylization with input content and style images. To transfer the style of an arbitrary image to a content image, these methods used a feed-forward network with a lowest-scaled feature transformer or a cascade of the networks with a feature transformer of a corresponding scale. However, their approaches did not consider either multi-scaled style in their single-scale feature transformer or dependency between the transformed feature statistics across the cascade networks. This shortcoming resulted in generating partially and inexactly transferred style in the generated images.
To overcome this limitation of partial style transfer, we propose a total style transferring method which transfers multi-scaled feature statistics through a single feed-forward process. First, our method transforms multi-scaled feature maps of a content image into those of a target style image by considering both inter-channel correlations in each single scaled feature map and inter-scale correlations between multi-scaled feature maps. Second, each transformed feature map is inserted into the decoder layer of the corresponding scale using skip-connection. Finally, the skip-connected multi-scaled feature maps are decoded into a stylized image through our trained decoder network. | 0reject
|
Title: AFINets: Attentive Feature Integration Networks for Image Classification. Abstract: Convolutional Neural Networks (CNNs) have achieved tremendous success in a number of learning tasks, e.g., image classification. Recent advances in CNNs, such as ResNets and DenseNets, mainly focus on the skip and concatenation operators to avoid gradient vanishing. However, such operators largely neglect information across layers (as in ResNets) or involve tremendous redundancy of features repeatedly copied from previous layers (as in DenseNets). In this paper, we design Attentive Feature Integration (AFI) modules, which can be applicable to most recent network architectures, leading to new architectures named as AFINets. AFINets can by adaptively integrate distinct information through explicitly modeling the subordinate relationship between different levels of features. Experimental results on benchmark datasets have demonstrated the effectiveness of the proposed AFI modules.
| 2withdrawn
|
Title: SINGLE PATH ONE-SHOT NEURAL ARCHITECTURE SEARCH WITH UNIFORM SAMPLING. Abstract: We revisit the one-shot Neural Architecture Search (NAS) paradigm and analyze its advantages over existing NAS approaches. Existing one-shot method (Benderet al., 2018), however, is hard to train and not yet effective on large scale datasets like ImageNet. This work propose a Single Path One-Shot model to address the challenge in the training. Our central idea is to construct a simplified supernet, where all architectures are single paths so that weight co-adaption problem is alleviated. Training is performed by uniform path sampling. All architectures (and their weights) are trained fully and equally.
Comprehensive experiments verify that our approach is flexible and effective. It is easy to train and fast to search. It effortlessly supports complex search spaces(e.g., building blocks, channel, mixed-precision quantization) and different search constraints (e.g., FLOPs, latency). It is thus convenient to use for various needs. It achieves start-of-the-art performance on the large dataset ImageNet. | 0reject
|
Title: AdaFocal: Calibration-aware Adaptive Focal Loss. Abstract: Much recent work has been devoted to the problem of ensuring that a neural network's confidence scores match the true probability of being correct, i.e. the calibration problem. Of note, it was found that training with Focal loss leads to better calibrated deep networks than cross-entropy loss, while achieving the same level of accuracy \cite{mukhoti2020}. This success stems from Focal loss regularizing the entropy of the network's prediction (controlled by the hyper-parameter $\gamma$), thereby reining in the network's overconfidence. Further improvements in calibration can be achieved if $\gamma$ is selected independently for each training sample. However, the proposed strategy (named FLSD-53) is based on simple heuristics which, when selecting the $\gamma$, does not take into account any knowledge of whether the network is under or over confident about such samples and by how much. As a result, in most cases, this strategy performs only slightly better. In this paper, we propose a calibration-aware sample-dependent Focal loss called AdaFocal that adaptively modifies $\gamma$ from one training step to the next based on the information about the network's current calibration behaviour. At each training step $t$, AdaFocal adjusts the $\gamma_t$ based on (1) $\gamma_{t-1}$ of the previous training step (2) the magnitude of the network's under/over-confidence. We evaluate our proposed method on various image recognition and NLP tasks, covering a variety of network architectures, and confirm that AdaFocal consistently achieves significantly better calibration than the competing state-of-the-art methods without loss of accuracy. | 0reject
|
Title: Proximal Gradient Descent-Ascent: Variable Convergence under KŁ Geometry. Abstract: The gradient descent-ascent (GDA) algorithm has been widely applied to solve minimax optimization problems. In order to achieve convergent policy parameters for minimax optimization, it is important that GDA generates convergent variable sequences rather than convergent sequences of function value or gradient norm. However, the variable convergence of GDA has been proved only under convexity geometries, and it is lack of understanding in general nonconvex minimax optimization. This paper fills such a gap by studying the convergence of a more general proximal-GDA for regularized nonconvex-strongly-concave minimax optimization. Specifically, we show that proximal-GDA admits a novel Lyapunov function, which monotonically decreases in the minimax optimization process and drives the variable sequences to a critical point. By leveraging this Lyapunov function and the KL geometry that parameterizes the local geometries of general nonconvex functions, we formally establish the variable convergence of proximal-GDA to a certain critical point $x^*$, i.e., $x_t\to x^*, y_t\to y^*(x^*)$. Furthermore, over the full spectrum of the KL-parameterized geometry, we show that proximal-GDA achieves different types of convergence rates ranging from sublinear convergence up to finite-step convergence, depending on the geometry associated with the KL parameter. This is the first theoretical result on the variable convergence for nonconvex minimax optimization. | 1accept
|
Title: Episodic Curiosity through Reachability. Abstract: Rewards are sparse in the real world and most of today's reinforcement learning algorithms struggle with such sparsity. One solution to this problem is to allow the agent to create rewards for itself - thus making rewards dense and more suitable for learning. In particular, inspired by curious behaviour in animals, observing something novel could be rewarded with a bonus. Such bonus is summed up with the real task reward - making it possible for RL algorithms to learn from the combined reward. We propose a new curiosity method which uses episodic memory to form the novelty bonus. To determine the bonus, the current observation is compared with the observations in memory. Crucially, the comparison is done based on how many environment steps it takes to reach the current observation from those in memory - which incorporates rich information about environment dynamics. This allows us to overcome the known "couch-potato" issues of prior work - when the agent finds a way to instantly gratify itself by exploiting actions which lead to hardly predictable consequences. We test our approach in visually rich 3D environments in ViZDoom, DMLab and MuJoCo. In navigational tasks from ViZDoom and DMLab, our agent outperforms the state-of-the-art curiosity method ICM. In MuJoCo, an ant equipped with our curiosity module learns locomotion out of the first-person-view curiosity only. The code is available at https://github.com/google-research/episodic-curiosity/. | 1accept
|
Title: Natural Posterior Network: Deep Bayesian Predictive Uncertainty for Exponential Family Distributions. Abstract: Uncertainty awareness is crucial to develop reliable machine learning models. In this work, we propose the Natural Posterior Network (NatPN) for fast and high-quality uncertainty estimation for any task where the target distribution belongs to the exponential family. Thus, NatPN finds application for both classification and general regression settings. Unlike many previous approaches, NatPN does not require out-of-distribution (OOD) data at training time. Instead, it leverages Normalizing Flows to fit a single density on a learned low-dimensional and task-dependent latent space. For any input sample, NatPN uses the predicted likelihood to perform a Bayesian update over the target distribution. Theoretically, NatPN assigns high uncertainty far away from training data. Empirically, our extensive experiments on calibration and OOD detection show that NatPN delivers highly competitive performance for classification, regression and count prediction tasks. | 1accept
|
Title: Accelerated Variance Reduced Stochastic Extragradient Method for Sparse Machine Learning Problems. Abstract: Recently, many stochastic gradient descent algorithms with variance reduction have been proposed. Moreover, their proximal variants such as Prox-SVRG can effectively solve non-smooth problems, which makes that they are widely applied in many machine learning problems. However, the introduction of proximal operator will result in the error of the optimal value. In order to address this issue, we introduce the idea of extragradient and propose a novel accelerated variance reduced stochastic extragradient descent (AVR-SExtraGD) algorithm, which inherits the advantages of Prox-SVRG and momentum acceleration techniques. Moreover, our theoretical analysis shows that AVR-SExtraGD enjoys the best-known convergence rates and oracle complexities of stochastic first-order algorithms such as Katyusha for both strongly convex and non-strongly convex problems. Finally, our experimental results show that for ERM problems and robust face recognition via sparse representation, our AVR-SExtraGD can yield the improved performance compared with Prox-SVRG and Katyusha. The asynchronous variant of AVR-SExtraGD outperforms KroMagnon and ASAGA, which are the asynchronous variants of SVRG and SAGA, respectively. | 0reject
|
Title: What Should Not Be Contrastive in Contrastive Learning. Abstract: Recent self-supervised contrastive methods have been able to produce impressive transferable visual representations by learning to be invariant to different data augmentations. However, these methods implicitly assume a particular set of representational invariances (e.g., invariance to color), and can perform poorly when a downstream task violates this assumption (e.g., distinguishing red vs. yellow cars). We introduce a contrastive learning framework which does not require prior knowledge of specific, task-dependent invariances. Our model learns to capture varying and invariant factors for visual representations by constructing separate embedding spaces, each of which is invariant to all but one augmentation. We use a multi-head network with a shared backbone which captures information across each augmentation and alone outperforms all baselines on downstream tasks. We further find that the concatenation of the invariant and varying spaces performs best across all tasks we investigate, including coarse-grained, fine-grained, and few-shot downstream classification tasks, and various data corruptions. | 1accept
|
Title: R-GSN: The Relation-based Graph Similar Network for Heterogeneous Graph. Abstract: Heterogeneous graph is a kind of data structure widely existing in real life. Nowadays, the research of graph neural networks on heterogeneous graphs has become more and more popular. The existing heterogeneous graph neural network algorithms mainly have two ideas, one is based on meta-path and the other is not. The idea based on meta-path often requires a lot of manual preprocessing, at the same time it is difficult to extend to large-scale graphs. In this paper, we proposed the general heterogeneous message passing paradigm and designed R-GSN that does not need meta-path, which is much improved compared to the baseline R-GCN. Experiments have shown that our R-GSN algorithm achieves the state-of-the-art performance on the ogbn-mag large-scale heterogeneous graph dataset. | 2withdrawn
|
Title: Towards Language Agnostic Universal Representations. Abstract: When a bilingual student learns to solve word problems in math, we expect the student to be able to solve these problem in both languages the student is fluent in, even if the math lessons were only taught in one language. However, current representations in machine learning are language dependent. In this work, we present a method to decouple the language from the problem by learning language agnostic representations and therefore allowing training a model in one language and applying to a different one in a zero shot fashion. We learn these representations by taking inspiration from linguistics, specifically the Universal Grammar hypothesis and learn universal latent representations that are language agnostic (Chomsky, 2014; Montague, 1970). We demonstrate the capabilities of these representations by showing that the models trained on a single language using language agnostic representations achieve very similar accuracies in other languages. | 0reject
|
Title: Learning Pseudometric-based Action Representations for Offline Reinforcement Learning. Abstract: Offline reinforcement learning is a promising approach for practical applications since it does not require interactions with real-world environments. However, existing offline RL methods only work well in environments with continuous or small discrete action spaces. In environments with large and discrete action spaces, such as recommender systems and dialogue systems, the performance of existing methods decreases drastically because they suffer from inaccurate value estimation for a large proportion of o.o.d. actions. While recent works have demonstrated that online RL benefits from incorporating semantic information in action representations, unfortunately, they fail to learn reasonable relative distances between action representations, which is key to offline RL to reduce the influence of out-of-distribution (o.o.d.) actions. This paper proposes an action representation learning framework for offline RL based on a pseudometric, which measures both the behavioral relation and the data-distributional relation between actions. We provide theoretical analysis on the continuity and the bounds of the expected Q-values using the learned action representations. Experimental results show that our methods significantly improve the performance of two typical offline RL methods in environments with large and discrete action spaces. | 0reject
|
Title: Pareto Frontier Approximation Network (PA-Net) Applied to Multi-objective TSP. Abstract: Multi-objective optimization is used in various areas of robotics like control, planning etc. Their solutions are dependent on multiple objective functions, which can be conflicting in nature. In such cases, the optimality is defined in terms of Pareto optimality. A set of these Pareto Optimal solutions in the objective space form a Pareto front (or frontier). Each solution has its own trade off. For instance, the travelling salesman problem (TSP) is used in robotics for task/resource allocation. Often this allocation is influenced by multiple objective functions and is solved using Multi-objective travelling salesman problem (MOTSP). In this work, we present PA-Net, a network that generates good approximations of the Pareto front for the multi-objective optimization problems. Our training framework is applicable to other multi-objective optimization problems; however, in this work, we focus on solving MOTSP. Firstly, MOTSP is converted into a constrained optimization problem. We then train our network to solve this constrained problem using the Lagrangian relaxation and policy gradient. With PA-Net we are able to generate better quality Pareto fronts with fast inference times as compared to other learning based and classical methods. Finally, we present the application of PA-Net to find optimal visiting order in coverage planning. | 2withdrawn
|
Title: Triangle and Four Cycle Counting with Predictions in Graph Streams. Abstract: We propose data-driven one-pass streaming algorithms for estimating the number of triangles and four cycles, two fundamental problems in graph analytics that are widely studied in the graph data stream literature. Recently, Hsu et al. (2019) and Jiang et al. (2020) applied machine learning techniques in other data stream problems, using a trained oracle that can predict certain properties of the stream elements to improve on prior “classical” algorithms that did not use oracles. In this paper, we explore the power of a “heavy edge” oracle in multiple graph edge streaming models. In the adjacency list model, we present a one-pass triangle counting algorithm improving upon the previous space upper bounds without such an oracle. In the arbitrary order model, we present algorithms for both triangle and four cycle estimation with fewer passes and the same space complexity as in previous algorithms, and we show several of these bounds are optimal. We analyze our algorithms under several noise models, showing that the algorithms perform well even when the oracle errs. Our methodology expands upon prior work on “classical” streaming algorithms, as previous multi-pass and random order streaming algorithms can be seen as special cases of our algorithms, where the first pass or random order was used to implement the heavy edge oracle. Lastly, our experiments demonstrate advantages of the proposed method compared to state-of-the-art streaming algorithms. | 1accept
|