arXiv subjects

Yang Song

Publications and source records attributed to Yang Song.

At least 307 records · Page 17Linked to original sources

arXiv2020

Autoregressive Score Matching

Autoregressive models use chain rule to define a joint probability distribution as a product of conditionals. These conditionals need to be normalized, imposing constraints on the functional families that can be used. To increase flexibility, we propose autoregressive conditional score models (AR-CSM) where we parameterize the joint distribution in terms of the derivatives of univariate log-conditionals (scores), which need not be normalized. To train AR-CSM, we introduce a new divergence between distributions named Composite Score Matching (CSM). For AR-CSM models, this divergence between data and model distributions can be computed and optimized efficiently, requiring no expensive sampling or adversarial training. Compared to previous score matching algorithms, our method is more scalable to high dimensional data and more stable to optimize. We show with extensive experimental results that it can be applied to density estimation on synthetic data, image generation, image denoising, and training latent variable models with implicit encoders.

cs.LG↗

arXiv2020

Improved Techniques for Training Score-Based Generative Models

Score-based generative models can produce high quality image samples comparable to GANs, without requiring adversarial optimization. However, existing training procedures are limited to images of low resolution (typically below 32x32), and can be unstable under some settings. We provide a new theoretical analysis of learning and sampling from score models in high dimensional spaces, explaining existing failure modes and motivating new solutions that generalize across datasets. To enhance stability, we also propose to maintain an exponential moving average of model weights. With these improvements, we can effortlessly scale score-based generative models to images with unprecedented resolutions ranging from 64x64 to 256x256. Our score-based models can generate high-fidelity samples that rival best-in-class GANs on various image datasets, including CelebA, FFHQ, and multiple LSUN categories.

cs.LG↗

arXiv2020

Imitation with Neural Density Models

We propose a new framework for Imitation Learning (IL) via density estimation of the expert's occupancy measure followed by Maximum Occupancy Entropy Reinforcement Learning (RL) using the density as a reward. Our approach maximizes a non-adversarial model-free RL objective that provably lower bounds reverse Kullback-Leibler divergence between occupancy measures of the expert and imitator. We present a practical IL algorithm, Neural Density Imitation (NDI), which obtains state-of-the-art demonstration efficiency on benchmark control tasks.

cs.LG↗

arXiv2020

Generative Modeling by Estimating Gradients of the Data Distribution

We introduce a new generative model where samples are produced via Langevin dynamics using gradients of the data distribution estimated with score matching. Because gradients can be ill-defined and hard to estimate when the data resides on low-dimensional manifolds, we perturb the data with different levels of Gaussian noise, and jointly estimate the corresponding scores, i.e., the vector fields of gradients of the perturbed data distribution for all noise levels. For sampling, we propose an annealed Langevin dynamics where we use gradients corresponding to gradually decreasing noise levels as the sampling process gets closer to the data manifold. Our framework allows flexible model architectures, requires no sampling during training or the use of adversarial methods, and provides a learning objective that can be used for principled model comparisons. Our models produce samples comparable to GANs on MNIST, CelebA and CIFAR-10 datasets, achieving a new state-of-the-art inception score of 8.87 on CIFAR-10. Additionally, we demonstrate that our models learn effective representations via image inpainting experiments.

cs.LG↗

arXiv2020

Understanding Classifier Mistakes with Generative Models

Although deep neural networks are effective on supervised learning tasks, they have been shown to be brittle. They are prone to overfitting on their training distribution and are easily fooled by small adversarial perturbations. In this paper, we leverage generative models to identify and characterize instances where classifiers fail to generalize. We propose a generative model of the features extracted by a classifier, and show using rigorous hypothesis testing that errors tend to occur when features are assigned low-probability by our model. From this observation, we develop a detection criteria for samples on which a classifier is likely to fail at test time. In particular, we test against three different sources of classification failures: mistakes made on the test set due to poor model generalization, adversarial samples and out-of-distribution samples. Our approach is agnostic to class labels from the training set which makes it applicable to models trained in a semi-supervised way.

cs.LG↗

arXiv2020

Learning to Match Jobs with Resumes from Sparse Interaction Data using Multi-View Co-Teaching Network

With the ever-increasing growth of online recruitment data, job-resume matching has become an important task to automatically match jobs with suitable resumes. This task is typically casted as a supervised text matching problem. Supervised learning is powerful when the labeled data is sufficient. However, on online recruitment platforms, job-resume interaction data is sparse and noisy, which affects the performance of job-resume match algorithms. To alleviate these problems, in this paper, we propose a novel multi-view co-teaching network from sparse interaction data for job-resume matching. Our network consists of two major components, namely text-based matching model and relation-based matching model. The two parts capture semantic compatibility in two different views, and complement each other. In order to address the challenges from sparse and noisy data, we design two specific strategies to combine the two components. First, two components share the learned parameters or representations, so that the original representations of each component can be enhanced. More importantly, we adopt a co-teaching mechanism to reduce the influence of noise in training data. The core idea is to let the two components help each other by selecting more reliable training instances. The two strategies focus on representation enhancement and data enhancement, respectively. Compared with pure text-based matching models, the proposed approach is able to learn better data representations from limited or even sparse interaction data, which is more resistible to noise in training data. Experiment results have demonstrated that our model is able to outperform state-of-the-art methods for job-resume matching.

cs.CL↗

arXiv2020

An affine reconstructed algorithm for diffusion on triangular grids using the nodal discontinuous Galerkin method

This work discusses the application of an affine reconstructed nodal DG method for unstructured grids of triangles. Solving the diffusion terms in the DG method is non-trivial due to the solution representations being piecewise continuous. Hence, the diffusive flux is not defined on the interface of elements. The proposed numerical approach reconstructs a smooth solution in a parallelogram that is enclosed by the quadrilateral formed by two adjacent triangle elements. The interface between these two triangles is the diagonal of the enclosed parallelogram. Similar to triangles, the mapping of parallelograms from a physical domain to a reference domain is an affine mapping, which is necessary for an accurate and efficient implementation of the numerical algorithm. Thus, all computations can still be performed on the reference domain, which promotes efficiency in computation and storage. This reconstruction does not make assumptions on choice of polynomial basis. Reconstructed DG algorithms have previously been developed for modal implementations of the convection-diffusion equations. However, to the best of the authors' knowledge, this is the first practical guideline that has been proposed for applying the reconstructed algorithm on a nodal discontinuous Galerkin method with a focus on accuracy and efficiency. The algorithm is demonstrated on a number of benchmark cases as well as a challenging substantive problem in HED hydrodynamics with highly disparate diffusion parameters.

physics.comp-ph↗

arXiv2020

Arbitrarily Strong Utility-Privacy Tradeoff in Multi-Agent Systems

Each agent in a network makes a local observation that is linearly related to a set of public and private parameters. The agents send their observations to a fusion center to allow it to estimate the public parameters. To prevent leakage of the private parameters, each agent first sanitizes its local observation using a local privacy mechanism before transmitting it to the fusion center. We investigate the utility-privacy tradeoff in terms of the Cramér-Rao lower bounds for estimating the public and private parameters. We study the class of privacy mechanisms given by linear compression and noise perturbation, and derive necessary and sufficient conditions for achieving arbitrarily strong utility-privacy tradeoff in a multi-agent system for both the cases where prior information is available and unavailable, respectively. We also provide a method to find the maximum estimation privacy achievable without compromising the utility and propose an alternating algorithm to optimize the utility-privacy tradeoff in the case where arbitrarily strong utility-privacy tradeoff is not achievable.

eess.SY↗

arXiv2020

Training Deep Energy-Based Models with f-Divergence Minimization

Deep energy-based models (EBMs) are very flexible in distribution parametrization but computationally challenging because of the intractable partition function. They are typically trained via maximum likelihood, using contrastive divergence to approximate the gradient of the KL divergence between data and model distribution. While KL divergence has many desirable properties, other f-divergences have shown advantages in training implicit density generative models such as generative adversarial networks. In this paper, we propose a general variational framework termed f-EBM to train EBMs using any desired f-divergence. We introduce a corresponding optimization algorithm and prove its local convergence property with non-linear dynamical systems theory. Experimental results demonstrate the superiority of f-EBM over contrastive divergence, as well as the benefits of training EBMs using f-divergences other than KL.

cs.LG↗

arXiv2020

Efficient Graph Generation with Graph Recurrent Attention Networks

We propose a new family of efficient and expressive deep generative models of graphs, called Graph Recurrent Attention Networks (GRANs). Our model generates graphs one block of nodes and associated edges at a time. The block size and sampling stride allow us to trade off sample quality for efficiency. Compared to previous RNN-based graph generative models, our framework better captures the auto-regressive conditioning between the already-generated and to-be-generated parts of the graph using Graph Neural Networks (GNNs) with attention. This not only reduces the dependency on node ordering but also bypasses the long-term bottleneck caused by the sequential nature of RNNs. Moreover, we parameterize the output distribution per block using a mixture of Bernoulli, which captures the correlations among generated edges within the block. Finally, we propose to handle node orderings in generation by marginalizing over a family of canonical orderings. On standard benchmarks, we achieve state-of-the-art time efficiency and sample quality compared to previous models. Additionally, we show our model is capable of generating large graphs of up to 5K nodes with good quality. To the best of our knowledge, GRAN is the first deep graph generative model that can scale to this size. Our code is released at: https://github.com/lrjconan/GRAN.

cs.LG↗

arXiv2020

Improving 3D Object Detection through Progressive Population Based Augmentation

Data augmentation has been widely adopted for object detection in 3D point clouds. However, all previous related efforts have focused on manually designing specific data augmentation methods for individual architectures. In this work, we present the first attempt to automate the design of data augmentation policies for 3D object detection. We introduce the Progressive Population Based Augmentation (PPBA) algorithm, which learns to optimize augmentation strategies by narrowing down the search space and adopting the best parameters discovered in previous iterations. On the KITTI 3D detection test set, PPBA improves the StarNet detector by substantial margins on the moderate difficulty category of cars, pedestrians, and cyclists, outperforming all current state-of-the-art single-stage detection models. Additional experiments on the Waymo Open Dataset indicate that PPBA continues to effectively improve the StarNet and PointPillars detectors on a 20x larger dataset compared to KITTI. The magnitude of the improvements may be comparable to advances in 3D perception architectures and the gains come without an incurred cost at inference time. In subsequent experiments, we find that PPBA may be up to 10x more data efficient than baseline 3D detection models without augmentation, highlighting that 3D detection models may achieve competitive accuracy with far fewer labeled examples.

cs.CV↗

arXiv2020

BiO-Net: Learning Recurrent Bi-directional Connections for Encoder-Decoder Architecture

U-Net has become one of the state-of-the-art deep learning-based approaches for modern computer vision tasks such as semantic segmentation, super resolution, image denoising, and inpainting. Previous extensions of U-Net have focused mainly on the modification of its existing building blocks or the development of new functional modules for performance gains. As a result, these variants usually lead to an unneglectable increase in model complexity. To tackle this issue in such U-Net variants, in this paper, we present a novel Bi-directional O-shape network (BiO-Net) that reuses the building blocks in a recurrent manner without introducing any extra parameters. Our proposed bi-directional skip connections can be directly adopted into any encoder-decoder architecture to further enhance its capabilities in various task domains. We evaluated our method on various medical image analysis tasks and the results show that our BiO-Net significantly outperforms the vanilla U-Net as well as other state-of-the-art methods. Our code is available at https://github.com/tiangexiang/BiO-Net.

cs.CV↗

arXiv2020

BrePartition: Optimized High-Dimensional kNN Search with Bregman Distances

Bregman distances (also known as Bregman divergences) are widely used in machine learning, speech recognition and signal processing, and kNN searches with Bregman distances have become increasingly important with the rapid advances of multimedia applications. Data in multimedia applications such as images and videos are commonly transformed into space of hundreds of dimensions. Such high-dimensional space has posed significant challenges for existing kNN search algorithms with Bregman distances, which could only handle data of medium dimensionality (typically less than 100). This paper addresses the urgent problem of high-dimensional kNN search with Bregman distances. We propose a novel partition-filter-refinement framework. Specifically, we propose an optimized dimensionality partitioning scheme to solve several non-trivial issues. First, an effective bound from each partitioned subspace to obtain exact kNN results is derived. Second, we conduct an in-depth analysis of the optimized number of partitions and devise an effective strategy for partitioning. Third, we design an efficient integrated index structure for all the subspaces together to accelerate the search processing. Moreover, we extend our exact solution to an approximate version by a trade-off between the accuracy and efficiency. Experimental results on four real-world datasets and two synthetic datasets show the clear advantage of our method in comparison to state-of-the-art algorithms.

cs.DB↗

arXiv2020

Detecting Problem Statements in Peer Assessments

Effective peer assessment requires students to be attentive to the deficiencies in the work they rate. Thus, their reviews should identify problems. But what ways are there to check that they do? We attempt to automate the process of deciding whether a review comment detects a problem. We use over 18,000 review comments that were labeled by the reviewees as either detecting or not detecting a problem with the work. We deploy several traditional machine-learning models, as well as neural-network models using GloVe and BERT embeddings. We find that the best performer is the Hierarchical Attention Network classifier, followed by the Bidirectional Gated Recurrent Units (GRU) Attention and Capsule model with scores of 93.1% and 90.5% respectively. The best non-neural network model was the support vector machine with a score of 89.71%. This is followed by the Stochastic Gradient Descent model and the Logistic Regression model with 89.70% and 88.98%.

cs.IR↗

arXiv2020

Gleason Score Prediction using Deep Learning in Tissue Microarray Image

Prostate cancer (PCa) is one of the most common cancers in men around the world. The most accurate method to evaluate lesion levels of PCa is microscopic inspection of stained biopsy tissue and estimate the Gleason score of tissue microarray (TMA) image by expert pathologists. However, it is time-consuming for pathologists to identify the cellular and glandular patterns for Gleason grading in large TMA images. We used Gleason2019 Challenge dataset to build a convolutional neural network (CNN) model to segment TMA images to regions of different Gleason grades and predict the Gleason score according to the grading segmentation. We used a pre-trained model of prostate segmentation to increase the accuracy of the Gleason grade segmentation. The model achieved a mean Dice of 75.6% on the test cohort and ranked 4th in the Gleason2019 Challenge with a score of 0.778 combined of Cohen's kappa and the f1-score.

eess.IV↗

arXiv2020

Unsupervised Instance Segmentation in Microscopy Images via Panoptic Domain Adaptation and Task Re-weighting

Unsupervised domain adaptation (UDA) for nuclei instance segmentation is important for digital pathology, as it alleviates the burden of labor-intensive annotation and domain shift across datasets. In this work, we propose a Cycle Consistency Panoptic Domain Adaptive Mask R-CNN (CyC-PDAM) architecture for unsupervised nuclei segmentation in histopathology images, by learning from fluorescence microscopy images. More specifically, we first propose a nuclei inpainting mechanism to remove the auxiliary generated objects in the synthesized images. Secondly, a semantic branch with a domain discriminator is designed to achieve panoptic-level domain adaptation. Thirdly, in order to avoid the influence of the source-biased features, we propose a task re-weighting mechanism to dynamically add trade-off weights for the task-specific loss functions. Experimental results on three datasets indicate that our proposed method outperforms state-of-the-art UDA methods significantly, and demonstrates a similar performance as fully supervised methods.

cs.CV↗

arXiv2020

Shape-Oriented Convolution Neural Network for Point Cloud Analysis

Point cloud is a principal data structure adopted for 3D geometric information encoding. Unlike other conventional visual data, such as images and videos, these irregular points describe the complex shape features of 3D objects, which makes shape feature learning an essential component of point cloud analysis. To this end, a shape-oriented message passing scheme dubbed ShapeConv is proposed to focus on the representation learning of the underlying shape formed by each local neighboring point. Despite this intra-shape relationship learning, ShapeConv is also designed to incorporate the contextual effects from the inter-shape relationship through capturing the long-ranged dependencies between local underlying shapes. This shape-oriented operator is stacked into our hierarchical learning architecture, namely Shape-Oriented Convolutional Neural Network (SOCNN), developed for point cloud analysis. Extensive experiments have been performed to evaluate its significance in the tasks of point cloud classification and part segmentation.

cs.CV↗

arXiv2020

Design of an MISO-SWIPT-Aided Code-Index Modulated Multi-Carrier M-DCSK System for e-Health IoT

Code index modulated multi-carrier M-ary differential chaos shift keying (CIM-MC-M-DCSK) system not only inherits low-power and low-complexity advantages of the conventional DCSK system, but also significantly increases the transmission rate. This feature is of particular importance to Internet of Things (IoT) with trillions of low-cost devices. In particular, for e-health IoT applications, an efficient transmission scheme is designed to solve the challenge of the limited battery capacity for numerous user equipments served by one base station. In this paper, a new multiple-input-single-output simultaneous wireless information and power transfer (MISO-SWIPT) scheme for CIM-MC-M-DCSK system is proposed by utilizing orthogonal characteristic of chaotic signals with different initial values. The proposed system adopts power splitting mode, which is very promising for simultaneously providing energy and transmitting information of the user equipments without any external power supply. In particular, the new system can achieve desirable anti-multipath-fading capability without using channel estimator. Moreover, the analytical bit-error-rate expression of the proposed system is derived over multipath Rayleigh fading channels. Furthermore, the spectral efficiency and energy efficiency of the proposed system are analyzed. Simulation results not only validate the analytical expressions, but also demonstrate the superiority of the proposed system.

cs.NI↗