Split (1)

train · 3.23k rows

source stringlengths 200 2.98k	target stringlengths 18 668
Memorization in over-parameterized neural networks can severely hurt generalization in the presence of mislabeled examples. However, mislabeled examples are to hard avoid in extremely large datasets. We address this problem using the implicit regularization effect of stochastic gradient descent with large learning rate...	We introduce a fast and easy-to-implement algorithm that is robust to dataset noise.
Binarized Neural Networks (BNNs) have recently attracted significant interest due to their computational efficiency. Concurrently, it has been shown that neural networks may be overly sensitive to ``attacks" -- tiny adversarial changes in the input -- which may be detrimental to their use in safety-critical domains. De...	Gradient-based attacks on binarized neural networks are not effective due to the non-differentiability of such networks; Our IPROP algorithm solves this problem using integer optimization
Highly regularized LSTMs achieve impressive results on several benchmark datasets in language modeling. We propose a new regularization method based on decoding the last token in the context using the predicted distribution of the next token. This biases the model towards retaining more contextual information, in turn ...	Decoding the last token in the context using the predicted next token distribution acts as a regularizer and improves language modeling.
The assumption that data samples are independently identically distributed is the backbone of many learning algorithms. Nevertheless, datasets often exhibit rich structures in practice, and we argue that there exist some unknown orders within the data instances. Aiming to find such orders, we introduce a novel Generati...	Propose to observe implicit orders in datasets in a generative model viewpoint.
We present a Neural Program Search, an algorithm to generate programs from natural language description and a small number of input / output examples. The algorithm combines methods from Deep Learning and Program Synthesis fields by designing rich domain-specific language (DSL) and defining efficient search algorithm g...	Program synthesis from natural language description and input / output examples via Tree-Beam Search over Seq2Tree model
Generative adversarial training can be generally understood as minimizing certain moment matching loss defined by a set of discriminator functions, typically neural networks. The discriminator set should be large enough to be able to uniquely identify the true distribution (discriminative), and also be small enough to...	This paper studies the discrimination and generalization properties of GANs when the discriminator set is a restricted function class like neural networks.
Normalization layers are a staple in state-of-the-art deep neural network architectures. They are widely believed to stabilize training, enable higher learning rate, accelerate convergence and improve generalization, though the reason for their effectiveness is still an active research topic. In this work, we challenge...	All you need to train deep residual networks is a good initialization; normalization layers are not necessary.
Designing a metric manually for unsupervised sequence generation tasks, such as text generation, is essentially difficult. In a such situation, learning a metric of a sequence from data is one possible solution. The previous study, SeqGAN, proposed the framework for unsupervised sequence generation, in which a metric i...	This paper aims to learn a better metric for unsupervised learning, such as text generation, and shows a significant improvement over SeqGAN.
One of the most successful techniques in generative models has been decomposing a complicated generation task into a series of simpler generation tasks. For example, generating an image at a low resolution and then learning to refine that into a high resolution image often improves results substantially. Here we ex...	Decompose the task of learning a generative model into learning disentangled latent factors for subsets of the data and then learning the joint over those latent factors.
Visual grounding of language is an active research field aiming at enriching text-based representations with visual information. In this paper, we propose a new way to leverage visual knowledge for sentence representations. Our approach transfers the structure of a visual representation space to the textual space by us...	We propose a joint model to incorporate visual knowledge in sentence representations
Mixed precision training (MPT) is becoming a practical technique to improve the speed and energy efficiency of training deep neural networks by leveraging the fast hardware support for IEEE half-precision floating point that is available in existing GPUs. MPT is typically used in combination with a technique called los...	We devise adaptive loss scaling to improve mixed precision training that surpass the state-of-the-art results.
Many real-world problems, e.g. object detection, have outputs that are naturally expressed as sets of entities. This creates a challenge for traditional deep neural networks which naturally deal with structured outputs such as vectors, matrices or tensors. We present a novel approach for learning to predict sets with u...	We present a novel approach for learning to predict sets with unknown permutation and cardinality using feed-forward deep neural networks.
Foveation is an important part of human vision, and a number of deep networks have also used foveation. However, there have been few systematic comparisons between foveating and non-foveating deep networks, and between different variable-resolution downsampling methods. Here we define several such methods, and compare ...	We compare object recognition performance on images that are downsampled uniformly and with three different foveation schemes.
We explore the concept of co-design in the context of neural network verification. Specifically, we aim to train deep neural networks that not only are robust to adversarial perturbations but also whose robustness can be verified more easily. To this end, we identify two properties of network models - weight sparsity a...	We develop methods to train deep neural models that are both robust to adversarial perturbations and whose robustness is significantly easier to verify.
Batch Normalization (BatchNorm) has shown to be effective for improving and accelerating the training of deep neural networks. However, recently it has been shown that it is also vulnerable to adversarial perturbations. In this work, we aim to investigate the cause of adversarial vulnerability of the BatchNorm. We hypo...	Investigation of how BatchNorm causes adversarial vulnerability and how to avoid it.
Electronic Health Records (EHR) comprise of longitudinal clinical observations portrayed with sparsity, irregularity, and high-dimensionality which become the major obstacles in drawing reliable downstream outcome. Despite greatly numbers of imputation methods are being proposed to tackle these issues, most of the exis...	Our variational-recurrent imputation network (V-RIN) takes into account the correlated features, temporal dynamics, and further utilizes the uncertainty to alleviate the risk of biased missing values estimates.
Despite the state-of-the-art accuracy of Deep Neural Networks (DNN) in various classification problems, their deployment onto resource constrained edge computing devices remains challenging due to their large size and complexity. Several recent studies have reported remarkable results in reducing this complexity throug...	An adaptive method for fixed-point quantization of neural networks based on theoretical analysis rather than heuristics.
We study the problem of learning permutation invariant representations that can capture containment relations. We propose training a model on a novel task: predicting the size of the symmetric difference between pairs of multisets, sets which may contain multiple copies of the same object. With motivation from fuzzy se...	Based on fuzzy set theory, we propose a model that given only the sizes of symmetric differences between pairs of multisets, learns representations of such multisets and their elements.
It is important to collect credible training samples $(x,y)$ for building data-intensive learning systems (e.g., a deep learning system). In the literature, there is a line of studies on eliciting distributional information from self-interested agents who hold a relevant information. Asking people to report complex d...	This paper proposes a deep learning aided method to elicit credible samples from self-interested agents.
The celebrated Sequence to Sequence learning (Seq2Seq) technique and its numerous variants achieve excellent performance on many tasks. However, many machine learning tasks have inputs naturally represented as graphs; existing Seq2Seq models face a significant challenge in achieving accurate conversion from graph form ...	Graph to Sequence Learning with Attention-Based Neural Networks
We address the problem of learning to discover 3D parts for objects in unseen categories. Being able to learn the geometry prior of parts and transfer this prior to unseen categories pose fundamental challenges on data-driven shape segmentation approaches. Formulated as a contextual bandit problem, we propose a learnin...	A zero-shot segmentation framework for 3D object part segmentation. Model the segmentation as a decision-making process and solve as a contextual bandit problem.
This paper presents the ballistic graph neural network. Ballistic graph neural network tackles the weight distribution from a transportation perspective and has many different properties comparing to the traditional graph neural network pipeline. The ballistic graph neural network does not require to calculate any eige...	A new perspective on how to collect the correlation between nodes based on diffusion properties.
In this paper, we propose a \textit{weak supervision} framework for neural ranking tasks based on the data programming paradigm \citep{Ratner2016}, which enables us to leverage multiple weak supervision signals from different sources. Empirically, we consider two sources of weak supervision signals, unsupervised rankin...	We propose a weak supervision training pipeline based on the data programming framework for ranking tasks, in which we train a BERT-base ranking model and establish the new SOTA.
We study the training process of Deep Neural Networks (DNNs) from the Fourier analysis perspective. We demonstrate a very universal Frequency Principle (F-Principle) --- DNNs often fit target functions from low to high frequencies --- on high-dimensional benchmark datasets, such as MNIST/CIFAR10, and deep networks, suc...	In real problems, we found that DNNs often fit target functions from low to high frequencies during the training process.
The problem of accelerating drug discovery relies heavily on automatic tools to optimize precursor molecules to afford them with better biochemical properties. Our work in this paper substantially extends prior state-of-the-art on graph-to-graph translation methods for molecular optimization. In particular, we realize ...	We propose a multi-resolution, hierarchically coupled encoder-decoder for graph-to-graph translation.
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 architect...	We utilize attention to restrict equivariant neural networks to the set or co-occurring transformations in data.
The fast generation and refinement of protein backbones would constitute a major advancement to current methodology for the design and development of de novo proteins. In this study, we train Generative Adversarial Networks (GANs) to generate fixed-length full-atom protein backbones, with the goal of sampling from the ...	We train a GAN to generate and recover full-atom protein backbones , and we show that in select cases we can recover the generated proteins after sequence design and ab initio forward-folding.
Few-Shot Learning (learning with limited labeled data) aims to overcome the limitations of traditional machine learning approaches which require thousands of labeled examples to train an effective model. Considered as a hallmark of human intelligence, the community has recently witnessed several contributions on this t...	Meta Learning for Few Shot learning assumes that training tasks and test tasks are drawn from the same distribution. What do you do if they are not? Meta Learning with task-level Domain Adaptation!
Universal probabilistic programming systems (PPSs) provide a powerful framework for specifying rich and complex probabilistic models. However, this expressiveness comes at the cost of substantially complicating the process of drawing inferences from the model. In particular, inference can become challenging when the su...	Divide, Conquer, and Combine is a new inference scheme that can be performed on the probabilistic programs with stochastic support, i.e. the very existence of variables is stochastic.
Detecting communities or the modular structure of real-life networks (e.g. a social network or a product purchase network) is an important task because the way a network functions is often determined by its communities. The traditional approaches to community detection involve modularity-based approaches, which gen...	A community preserving node embedding algorithm that results in more effective detection of communities with a clustering on the embedded space
A point cloud is an agile 3D representation, efficiently modeling an object's surface geometry. However, these surface-centric properties also pose challenges on designing tools to recognize and synthesize point clouds. This work presents a novel autoregressive model, PointGrow, which generates realistic point cloud sa...	An autoregressive deep learning model for generating diverse point clouds.
Reinforcement learning and evolutionary algorithms can be used to create sophisticated control solutions. Unfortunately explaining how these solutions work can be difficult to due to their "black box" nature. In addition, the time-extended nature of control algorithms often prevent direct applications of explainability...	Describes a series of explainability techniques applied to a simple neural network controller used for navigation.
The Vision-and-Language Navigation (VLN) task entails an agent following navigational instruction in photo-realistic unknown environments. This challenging task demands that the agent be aware of which instruction was completed, which instruction is needed next, which way to go, and its navigation progress towards the ...	We propose a self-monitoring agent for the Vision-and-Language Navigation task.
Environments in Reinforcement Learning (RL) are usually only partially observable. To address this problem, a possible solution is to provide the agent with information about past observations. While common methods represent this history using a Recurrent Neural Network (RNN), in this paper we propose an alternative r...	event discovery to represent the history for the agent in RL
The unconditional generation of high fidelity images is a longstanding benchmark for testing the performance of image decoders. Autoregressive image models have been able to generate small images unconditionally, but the extension of these methods to large images where fidelity can be more readily assessed has rema...	We show that autoregressive models can generate high fidelity images.
Real-world dynamical systems often consist of multiple stochastic subsystems that interact with each other. Modeling and forecasting the behavior of such dynamics are generally not easy, due to the inherent hardness in understanding the complicated interactions and evolutions of their constituents. This paper introduce...	A deep hierarchical state-space model in which the state transitions of correlated objects are coordinated by graph neural networks.
Natural language is hierarchically structured: smaller units (e.g., phrases) are nested within larger units (e.g., clauses). When a larger constituent ends, all of the smaller constituents that are nested within it must also be closed. While the standard LSTM architecture allows different neurons to track information a...	We introduce a new inductive bias that integrates tree structures in recurrent neural networks.
Skip connections made the training of very deep networks possible and have become an indispensable component in a variety of neural architectures. A completely satisfactory explanation for their success remains elusive. Here, we present a novel explanation for the benefits of skip connections in training very deep netw...	Degenerate manifolds arising from the non-identifiability of the model slow down learning in deep networks; skip connections help by breaking degeneracies.
Representation learning is a central challenge across a range of machine learning areas. In reinforcement learning, effective and functional representations have the potential to tremendously accelerate learning progress and solve more challenging problems. Most prior work on representation learning has focused on gene...	Learning state representations which capture factors necessary for control
We explore the behavior of a standard convolutional neural net in a setting that introduces classification tasks sequentially and requires the net to master new tasks while preserving mastery of previously learned tasks. This setting corresponds to that which human learners face as they acquire domain expertise, for ...	We study the behavior of a CNN as it masters new tasks while preserving mastery for previously learned tasks
We demonstrate a low effort method that unsupervisedly constructs task-optimized embeddings from existing word embeddings to gain performance on a supervised end-task. This avoids additional labeling or building more complex model architectures by instead providing specialized embeddings better fit for the end-task(s)....	Morty refits pretrained word embeddings to either: (a) improve overall embedding performance (for Multi-task settings) or improve Single-task performance, while requiring only minimal effort.
Data augmentation is commonly used to encode invariances in learning methods. However, this process is often performed in an inefficient manner, as artificial examples are created by applying a number of transformations to all points in the training set. The resulting explosion of the dataset size can be an issue in te...	Selectively augmenting difficult to classify points results in efficient training.
Over the last few years exciting work in deep generative models has produced models able to suggest new organic molecules by generating strings, trees, and graphs representing their structure. While such models are able to generate molecules with desirable properties, their utility in practice is limited due to the dif...	A deep generative model for organic molecules that first generates reactant building blocks before combining these using a reaction predictor.
Deep neural networks are complex non-linear models used as predictive analytics tool and have demonstrated state-of-the-art performance on many classification tasks. However, they have no inherent capability to recognize when their predictions might go wrong. There have been several efforts in the recent past to dete...	Improve the robustness and energy efficiency of a deep neural network using the hidden representations.
Many methods have been developed to represent knowledge graph data, which implicitly exploit low-rank latent structure in the data to encode known information and enable unknown facts to be inferred. To predict whether a relationship holds between entities, their embeddings are typically compared in the latent space fo...	Understanding the structure of knowledge graph representation using insight from word embeddings.
Many real-world applications involve multivariate, geo-tagged time series data: at each location, multiple sensors record corresponding measurements. For example, air quality monitoring system records PM2.5, CO, etc. The resulting time-series data often has missing values due to device outages or communication errors. ...	A novel self-attention mechanism for multivariate, geo-tagged time series imputation.
The conversion of scanned documents to digital forms is performed using an Optical Character Recognition (OCR) software. This work focuses on improving the quality of scanned documents in order to improve the OCR output. We create an end-to-end document enhancement pipeline which takes in a set of noisy documents and p...	We designed and tested a REDNET (ResNet Encoder-Decoder) with 8 skip connections to remove noise from documents, including blurring and watermarks, resulting in a high performance deep network for document image cleanup.
The existence of adversarial examples, or intentional mis-predictions constructed from small changes to correctly predicted examples, is one of the most significant challenges in neural network research today. Ironically, many new defenses are based on a simple observation - the adversarial inputs themselves are not ro...	We identify a family of defense techniques and show that both deterministic lossy compression and randomized perturbations to the input lead to similar gains in robustness.
There is no consensus yet on the question whether adaptive gradient methods like Adam are easier to use than non-adaptive optimization methods like SGD. In this work, we fill in the important, yet ambiguous concept of ‘ease-of-use’ by defining an optimizer’s tunability: How easy is it to find good hyperparameter confi...	We provide a method to benchmark optimizers that is cognizant to the hyperparameter tuning process.
The phase problem in diffraction physics is one of the oldest inverse problems in all of science. The central difficulty that any approach to solving this inverse problem must overcome is that half of the information, namely the phase of the diffracted beam, is always missing. In the context of electron microscopy, the...	We introduce a semi-supervised deep neural network to approximate the solution of the phase problem in electron microscopy
Word embeddings extract semantic features of words from large datasets of text. Most embedding methods rely on a log-bilinear model to predict the occurrence of a word in a context of other words. Here we propose word2net, a method that replaces their linear parametrization with neural networks. For each term in the...	Word2net is a novel method for learning neural network representations of words that can use syntactic information to learn better semantic features.
A key goal in neuroscience is to understand brain mechanisms of cognitive functions. An emerging approach is to study “brain states” dynamics using functional magnetic resonance imaging (fMRI). So far in the literature, brain states have typically been studied using 30 seconds of fMRI data or more, and it is unclear to...	Using a 10s window of fMRI signals, our GCN model identified 21 different task conditions from HCP dataset with a test accuracy of 89%.
Modern deep neural networks (DNNs) require high memory consumption and large computational loads. In order to deploy DNN algorithms efficiently on edge or mobile devices, a series of DNN compression algorithms have been explored, including the line of works on factorization methods. Factorization methods approximate ...	Efficiently inducing low-rank deep neural networks via SVD training with sparse singular values and orthogonal singular vectors.
The recent rise in popularity of few-shot learning algorithms has enabled models to quickly adapt to new tasks based on only a few training samples. Previous few-shot learning works have mainly focused on classification and reinforcement learning. In this paper, we propose a few-shot meta-learning system that focuses...	We propose a few-shot learning model that is tailored specifically for regression tasks
Most classification and segmentation datasets assume a closed-world scenario in which predictions are expressed as distribution over a predetermined set of visual classes. However, such assumption implies unavoidable and often unnoticeable failures in presence of out-of-distribution (OOD) input. These failures are boun...	We present a novel approach for detecting out-of-distribution pixels in semantic segmentation.
Network quantization is one of the most hardware friendly techniques to enable the deployment of convolutional neural networks (CNNs) on low-power mobile devices. Recent network quantization techniques quantize each weight kernel in a convolutional layer independently for higher inference accuracy, since the weight ker...	Accurate, Fast and Automated Kernel-Wise Neural Network Quantization with Mixed Precision using Hierarchical Deep Reinforcement Learning
Recent visual analytics systems make use of multiple machine learning models to better fit the data as opposed to traditional single, pre-defined model systems. However, while multi-model visual analytic systems can be effective, their added complexity poses usability concerns, as users are required to interact with th...	Gaggle, an interactive visual analytic system to help users interactively navigate model space for classification and ranking tasks.
Chinese text classification has received more and more attention today. However, the problem of Chinese text representation still hinders the improvement of Chinese text classification, especially the polyphone and the homophone in social media. To cope with it effectively, we propose a new structure, the Extractor, ba...	We propose a novel attention networks with the hybird encoder to solve the text representation issue of Chinese text classification, especially the language phenomena about pronunciations such as the polyphone and the homophone.
Recent advances in learning from demonstrations (LfD) with deep neural networks have enabled learning complex robot skills that involve high dimensional perception such as raw image inputs. LfD algorithms generally assume learning from single task demonstrations. In practice, however, it is more efficient for a teach...	multi-modal imitation learning from unstructured demonstrations using stochastic neural network modeling intention.
The interpretability of neural networks has become crucial for their applications in real world with respect to the reliability and trustworthiness. Existing explanation generation methods usually provide important features by scoring their individual contributions to the model prediction and ignore the interactions be...	A novel approach to construct hierarchical explanations for text classification by detecting feature interactions.
Making deep convolutional neural networks more accurate typically comes at the cost of increased computational and memory resources. In this paper, we reduce this cost by exploiting the fact that the importance of features computed by convolutional layers is highly input-dependent, and propose feature boosting and supp...	We make convolutional layers run faster by dynamically boosting and suppressing channels in feature computation.
We propose a novel way of reducing the number of parameters in the storage-hungry fully connected layers of a neural network by using pre-defined sparsity, where the majority of connections are absent prior to starting training. Our results indicate that convolutional neural networks can operate without any loss of acc...	Neural networks can be pre-defined to have sparse connectivity without performance degradation.
Deep neural networks are vulnerable to adversarial examples, which becomes one of the most important problems in the development of deep learning. While a lot of efforts have been made in recent years, it is of great significance to perform correct and complete evaluations of the adversarial attack and defense algorith...	We provide a comprehensive, rigorous, and coherent benchmark to evaluate adversarial robustness of deep learning models.
We propose a modification to traditional Artificial Neural Networks (ANNs), which provides the ANNs with new aptitudes motivated by biological neurons. Biological neurons work far beyond linearly summing up synaptic inputs and then transforming the integrated information. A biological neuron change firing modes acc...	We propose a modification to traditional Artificial Neural Networks motivated by the biology of neurons to enable the shape of the activation function to be context dependent.
In this work, we study how the large-scale pretrain-finetune framework changes the behavior of a neural language generator. We focus on the transformer encoder-decoder model for the open-domain dialogue response generation task. We find that after standard fine-tuning, the model forgets important language generation sk...	We identify the forgetting problem in fine-tuning of pre-trained NLG models, and propose the mix-review strategy to address it.
Combining domain knowledge models with neural models has been challenging. End-to-end trained neural models often perform better (lower Mean Square Error) than domain knowledge models or domain/neural combinations, and the combination is inefficient to train. In this paper, we demonstrate that by composing domain m...	Improved modeling of complex systems uses hybrid neural/domain model composition, new decorrelation loss functions and extrapolative test sets
Humans can learn task-agnostic priors from interactive experience and utilize the priors for novel tasks without any finetuning. In this paper, we propose Scoring-Aggregating-Planning (SAP), a framework that can learn task-agnostic semantics and dynamics priors from arbitrary quality interactions as well as the corresp...	We learn dense scores and dynamics model as priors from exploration data and use them to induce a good policy in new tasks in zero-shot condition.
Particle-based inference algorithm is a promising method to efficiently generate samples for an intractable target distribution by iteratively updating a set of particles. As a noticeable example, Stein variational gradient descent (SVGD) provides a deterministic and computationally efficient update, but it is known to...	Analyze the underlying mechanisms of variance collapse of SVGD in high dimensions.
We describe an approach to understand the peculiar and counterintuitive generalization properties of deep neural networks. The approach involves going beyond worst-case theoretical capacity control frameworks that have been popular in machine learning in recent years to revisit old ideas in the statistical mechanics ...	Rethinking generalization requires revisiting old ideas: statistical mechanics approaches and complex learning behavior
Computations for the softmax function in neural network models are expensive when the number of output classes is large. This can become a significant issue in both training and inference for such models. In this paper, we present Doubly Sparse Softmax (DS-Softmax), Sparse Mixture of Sparse of Sparse Experts, to improv...	We present doubly sparse softmax, the sparse mixture of sparse of sparse experts, to improve the efficiency for softmax inference through exploiting the two-level overlapping hierarchy.
Supervised machine learning models for high-value computer vision applications such as medical image classification often require large datasets labeled by domain experts, which are slow to collect, expensive to maintain, and static with respect to changes in the data distribution. In this context, we assess the utilit...	We explore using passively collected eye-tracking data to reduce the amount of labeled data needed during training.
We study the robustness to symmetric label noise of GNNs training procedures. By combining the nonlinear neural message-passing models (e.g. Graph Isomorphism Networks, GraphSAGE, etc.) with loss correction methods, we present a noise-tolerant approach for the graph classification task. Our experiments show that test a...	We apply loss correction to graph neural networks to train a more robust to noise model.
Through many recent advances in graph representation learning, performance achieved on tasks involving graph-structured data has substantially increased in recent years---mostly on tasks involving node-level predictions. The setup of prediction tasks over entire graphs (such as property prediction for a molecule, or si...	We use graph co-attention in a paired graph training system for graph classification and regression.
In this paper we study image captioning as a conditional GAN training, proposing both a context-aware LSTM captioner and co-attentive discriminator, which enforces semantic alignment between images and captions. We investigate the viability of two discrete GAN training methods: Self-critical Sequence Training (SCST) an...	Image captioning as a conditional GAN training with novel architectures, also study two discrete GAN training methods.
We present Newtonian Monte Carlo (NMC), a method to improve Markov Chain Monte Carlo (MCMC) convergence by analyzing the first and second order gradients of the target density to determine a suitable proposal density at each point. Existing first order gradient-based methods suffer from the problem of determining an ap...	Exploit curvature to make MCMC methods converge faster than state of the art.
Neural Tangents is a library designed to enable research into infinite-width neural networks. It provides a high-level API for specifying complex and hierarchical neural network architectures. These networks can then be trained and evaluated either at finite-width as usual or in their infinite-width limit. Infinite-wid...	Keras for infinite neural networks.
Deep neural networks have achieved great success in classiﬁcation tasks during the last years. However, one major problem to the path towards artiﬁcial intelligence is the inability of neural networks to accurately detect samples from novel class distributions and therefore, most of the existent classiﬁcation algorithm...	We propose a novel loss function that achieves state-of-the-art results in out-of-distribution detection with Outlier Exposure both on image and text classiﬁcation tasks.
Navigation is crucial for animal behavior and is assumed to require an internal representation of the external environment, termed a cognitive map. The precise form of this representation is often considered to be a metric representation of space. An internal representation, however, is judged by its contribution to pe...	Task agnostic pre-training can shape RNN's attractor landscape, and form diverse inductive bias for different navigation tasks
Formal verification of machine learning models has attracted attention recently, and significant progress has been made on proving simple properties like robustness to small perturbations of the input features. In this context, it has also been observed that folding the verification procedure into training makes it eas...	Neural Network Verification for Temporal Properties and Sequence Generation Models
Neural Network (NN) has achieved state-of-the-art performances in many tasks within image, speech, and text domains. Such great success is mainly due to special structure design to fit the particular data patterns, such as CNN capturing spatial locality and RNN modeling sequential dependency. Essentially, these specifi...	We propose a universal neural network solution to derive effective NN architectures for tabular data automatically.
Knowledge Bases (KBs) are becoming increasingly large, sparse and probabilistic. These KBs are typically used to perform query inferences and rule mining. But their efficacy is only as high as their completeness. Efficiently utilizing incomplete KBs remains a major challenge as the current KB completion techniques eith...	Probabilistic Rule Learning system using Lifted Inference
Recent efforts in Dialogue State Tracking (DST) for task-oriented dialogues have progressed toward open-vocabulary or generation-based approaches where the models can generate slot value candidates from the dialogue history itself. These approaches have shown good performance gain, especially in complicated dialogue do...	We propose the first non-autoregressive neural model for Dialogue State Tracking (DST), achieving the SOTA accuracy (49.04%) on MultiWOZ2.1 benchmark, and reducing inference latency by an order of magnitude.
The 3D-zoom operation is the positive translation of the camera in the Z-axis, perpendicular to the image plane. In contrast, the optical zoom changes the focal length and the digital zoom is used to enlarge a certain region of an image to the original image size. In this paper, we are the first to formulate an unsuper...	A novel network architecture to perform Deep 3D Zoom or close-ups.
The universal approximation theorem, in one of its most general versions, says that if we consider only continuous activation functions σ, then a standard feedforward neural network with one hidden layer is able to approximate any continuous multivariate function f to any given approximation threshold ε, if and only if...	A quantitative refinement of the universal approximation theorem via an algebraic approach.
In this paper, we design a generic framework for learning a robust text classification model that achieves accuracy comparable to standard full models under test-time budget constraints. We take a different approach from existing methods and learn to dynamically delete a large fraction of unimportant words by a low-co...	Modular framework for document classification and data aggregation technique for making the framework robust to various distortion, and noise and focus only on the important words.
Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-net and searching for an optimal architecture. In this paper, we present a novel approach, namely Partially-Connected DARTS,...	Allowing partial channel connection in super-networks to regularize and accelerate differentiable architecture search
Dialogue research tends to distinguish between chit-chat and goal-oriented tasks. While the former is arguably more naturalistic and has a wider use of language, the latter has clearer metrics and a more straightforward learning signal. Humans effortlessly combine the two, and engage in chit-chat for example with the g...	Agents interact (speak, act) and can achieve goals in a rich world with diverse language, bridging the gap between chit-chat and goal-oriented dialogue.
We consider off-policy policy evaluation when the trajectory data are generated by multiple behavior policies. Recent work has shown the key role played by the state or state-action stationary distribution corrections in the infinite horizon context for off-policy policy evaluation. We propose estimated mixture policy ...	A new partially policy-agnostic method for infinite-horizon off-policy policy evalution with multiple known or unknown behavior policies.
We introduce a more efficient neural architecture for amortized inference, which combines continuous and conditional normalizing flows using a principled choice of structure. Our gradient flow derives its sparsity pattern from the minimally faithful inverse of its underlying graphical model. We find that this factoriza...	We introduce a more efficient neural architecture for amortized inference, which combines continuous and conditional normalizing flows using a principled choice of sparsity structure.
We present a neural architecture search algorithm to construct compact reinforcement learning (RL) policies, by combining ENAS and ES in a highly scalable and intuitive way. By defining the combinatorial search space of NAS to be the set of different edge-partitionings (colorings) into same-weight classes, we represent...	We show that ENAS with ES-optimization in RL is highly scalable, and use it to compactify neural network policies by weight sharing.
Deep approaches to anomaly detection have recently shown promising results over shallow methods on large and complex datasets. Typically anomaly detection is treated as an unsupervised learning problem. In practice however, one may have---in addition to a large set of unlabeled samples---access to a small pool of label...	We introduce Deep SAD, a deep method for general semi-supervised anomaly detection that especially takes advantage of labeled anomalies.
To analyze deep ReLU network, we adopt a student-teacher setting in which an over-parameterized student network learns from the output of a fixed teacher network of the same depth, with Stochastic Gradient Descent (SGD). Our contributions are two-fold. First, we prove that when the gradient is zero (or bounded above by...	This paper analyzes training dynamics and critical points of training deep ReLU network via SGD in the teacher-student setting.
We study the convergence of gradient descent (GD) and stochastic gradient descent (SGD) for training $L$-hidden-layer linear residual networks (ResNets). We prove that for training deep residual networks with certain linear transformations at input and output layers, which are fixed throughout training, both GD and SGD...	Under certain condition on the input and output linear transformations, both GD and SGD can achieve global convergence for training deep linear ResNets.
In this paper, we empirically investigate the training journey of deep neural networks relative to fully trained shallow machine learning models. We observe that the deep neural networks (DNNs) train by learning to correctly classify shallow-learnable examples in the early epochs before learning the harder examples. We...	We analyze the training process for Deep Networks and show that they start from rapidly learning shallow classifiable examples and slowly generalize to harder data points.
While much recent work has targeted learning deep discrete latent variable models with variational inference, this setting remains challenging, and it is often necessary to make use of potentially high-variance gradient estimators in optimizing the ELBO. As an alternative, we propose to optimize a non-ELBO objective de...	Learning deep latent variable MRFs with a saddle-point objective derived from the Bethe partition function approximation.
In an explanation generation problem, an agent needs to identify and explain the reasons for its decisions to another agent. Existing work in this area is mostly confined to planning-based systems that use automated planning approaches to solve the problem. In this paper, we approach this problem from a new perspective...	A general framework for explanation generation using Logic.
Recent theoretical work has demonstrated that deep neural networks have superior performance over shallow networks, but their training is more difficult, e.g., they suffer from the vanishing gradient problem. This problem can be typically resolved by the rectified linear unit (ReLU) activation. However, here we show th...	Deep and narrow neural networks will converge to erroneous mean or median states of the target function depending on the loss with high probability.
We study adversarial robustness of neural networks from a margin maximization perspective, where margins are defined as the distances from inputs to a classifier's decision boundary. Our study shows that maximizing margins can be achieved by minimizing the adversarial loss on the decision boundary at the "shortest suc...	We propose MMA training to directly maximize input space margin in order to improve adversarial robustness primarily by removing the requirement of specifying a fixed distortion bound.
Many anomaly detection methods exist that perform well on low-dimensional problems however there is a notable lack of effective methods for high-dimensional spaces, such as images. Inspired by recent successes in deep learning we propose a novel approach to anomaly detection using generative adversarial networks. Given...	We propose a method for anomaly detection with GANs by searching the generator's latent space for good sample representations.
Variational inference (VI) and Markov chain Monte Carlo (MCMC) are approximate posterior inference algorithms that are often said to have complementary strengths, with VI being fast but biased and MCMC being slower but asymptotically unbiased. In this paper, we analyze gradient-based MCMC and VI procedures and find the...	The transient behavior of gradient-based MCMC and variational inference algorithms is more similar than one might think, calling into question the claim that variational inference is faster than MCMC.

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.