arXiv subjects

Xin Chen

Publications and source records attributed to Xin Chen.

At least 631 records · Page 35Linked to original sources

arXiv2018

Unsupervised Domain Adaptation using Generative Models and Self-ensembling

Transferring knowledge across different datasets is an important approach to successfully train deep models with a small-scale target dataset or when few labeled instances are available. In this paper, we aim at developing a model that can generalize across multiple domain shifts, so that this model can adapt from a single source to multiple targets. This can be achieved by randomizing the generation of the data of various styles to mitigate the domain mismatch. First, we present a new adaptation to the CycleGAN model to produce stochastic style transfer between two image batches of different domains. Second, we enhance the classifier performance by using a self-ensembling technique with a teacher and student model to train on both original and generated data. Finally, we present experimental results on three datasets Office-31, Office-Home, and Visual Domain adaptation. The results suggest that selfensembling is better than simple data augmentation with the newly generated data and a single model trained this way can have the best performance across all different transfer tasks.

cs.CV↗

arXiv2018

Beyond "How may I help you?": Assisting Customer Service Agents with Proactive Responses

We study the problem of providing recommended responses to customer service agents in live-chat dialogue systems. Smart-reply systems have been widely applied in real-world applications (e.g. Gmail, LinkedIn Messaging), where most of them can successfully recommend reactive responses. However, we observe a major limitation of current methods is that they generally have difficulties in suggesting proactive investigation act (e.g. "Do you perhaps have another account with us?") due to the lack of long-term context information, which indeed act as critical steps for customer service agents to collect information and resolve customers' issues. Thus in this work, we propose an end-to-end method with special focus on suggesting proactive investigative questions to customer agents in Airbnb's customer service live-chat system. Effectiveness of our proposed method can be validated through qualitative and quantitative results.

cs.IR↗

arXiv2018

Heat kernel for non-local operators with variable order

Let $α(x)$ be a measurable function taking values in $ [α_1,α_2]$ for $0<\A_1\le \A_2<2$, and $κ(x,z)$ be a positive measurable function that is symmetric in $z$ and bounded between two positive constants. Under a uniform Hölder continuous assumptions on $α(x)$ and $x\mapsto κ(x,z)$, we obtain existence, upper and lower bounds, and regularity properties of the heat kernel associated with the following non-local operator of variable order $$ \LL f(x)=\int_{\R^d}\big(f(x+z)-f(x)-\langle\nabla f(x), z\rangle \I_{\{|z|\le 1\}}\big) \frac{κ(x,z)}{|z|^{d+α(x)}}\,dz. $$ In particular, we show that the operator $\LL$ generates a conservative Feller process on $\R^d$ having the strong Feller property, which is usually assumed a priori in the literature to study analytic properties of $\LL$ via probabilistic approaches. Our near-diagonal estimates and lower bound estimates of the heat kernel depend on the local behavior of index function $α(x)$, when $α(x)\equiv \A\in(0,2)$, our results recover some results by Chen and Kumagai (2003) and Chen and Zhang (2016).

math.PR↗

arXiv2018

Bridging Semantic Gaps between Natural Languages and APIs with Word Embedding

Developers increasingly rely on text matching tools to analyze the relation between natural language words and APIs. However, semantic gaps, namely textual mismatches between words and APIs, negatively affect these tools. Previous studies have transformed words or APIs into low-dimensional vectors for matching; however, inaccurate results were obtained due to the failure of modeling words and APIs simultaneously. To resolve this problem, two main challenges are to be addressed: the acquisition of massive words and APIs for mining and the alignment of words and APIs for modeling. Therefore, this study proposes Word2API to effectively estimate relatedness of words and APIs. Word2API collects millions of commonly used words and APIs from code repositories to address the acquisition challenge. Then, a shuffling strategy is used to transform related words and APIs into tuples to address the alignment challenge. Using these tuples, Word2API models words and APIs simultaneously. Word2API outperforms baselines by 10%-49.6% of relatedness estimation in terms of precision and NDCG. Word2API is also effective on solving typical software tasks, e.g., query expansion and API documents linking. A simple system with Word2API-expanded queries recommends up to 21.4% more related APIs for developers. Meanwhile, Word2API improves comparison algorithms by 7.9%-17.4% in linking questions in Question&Answer communities to API documents.

cs.SE↗

arXiv2018

2PFPCE: Two-Phase Filter Pruning Based on Conditional Entropy

Deep Convolutional Neural Networks~(CNNs) offer remarkable performance of classifications and regressions in many high-dimensional problems and have been widely utilized in real-word cognitive applications. However, high computational cost of CNNs greatly hinder their deployment in resource-constrained applications, real-time systems and edge computing platforms. To overcome this challenge, we propose a novel filter-pruning framework, two-phase filter pruning based on conditional entropy, namely \textit{2PFPCE}, to compress the CNN models and reduce the inference time with marginal performance degradation. In our proposed method, we formulate filter pruning process as an optimization problem and propose a novel filter selection criteria measured by conditional entropy. Based on the assumption that the representation of neurons shall be evenly distributed, we also develop a maximum-entropy filter freeze technique that can reduce over fitting. Two filter pruning strategies -- global and layer-wise strategies, are compared. Our experiment result shows that combining these two strategies can achieve a higher neural network compression ratio than applying only one of them under the same accuracy drop threshold. Two-phase pruning, that is, combining both global and layer-wise strategies, achieves 10 X FLOPs reduction and 46% inference time reduction on VGG-16, with 2% accuracy drop.

cs.CV↗

arXiv2018

Stochastic variational principles for dissipative equations with advected quantities

This paper presents symmetry reduction for material stochastic Lagrangian systems with advected quantities whose configuration space is a Lie group. Such variational principles yield deterministic as well as stochastic constrained variational principles for dissipative equations of motion in spatial representation. The general theory is presented for the finite dimensional situation. In infinite dimensions we obtain partial differential equations and stochastic partial differential equations. When the Lie group is, for example, a diffeomorphism group, the general result is not directly applicable but the setup and method suggest rigorous proofs valid in infinite dimensions which lead to similar results. We apply this technique to the compressible Navier-Stokes equation and to magnetohydrodynamics for charged viscous compressible fluids. A stochastic Kelvin-Noether theorem is presented. We derive, among others, the classical deterministic dissipative equations from purely variational and stochastic principles, without any appeal to thermodynamics.

math-ph↗

arXiv2018

Random conductance models with stable-like jumps: heat kernel estimates and Harnack inequalities

We establish two-sided heat kernel estimates for random conductance models with non-uniformly elliptic (possibly degenerate) stable-like jumps on graphs. These are long range counterparts of well known two-sided Gaussian heat kernel estimates by M.T. Barlow for nearest neighbor (short range) random walks on the supercritical percolation cluster. Unlike the cases for nearest neighbor conductance models, the idea through parabolic Harnack inequalities does not work, since even elliptic Harnack inequalities do not hold in the present setting. As an application, we establish the local limit theorem for the models.

math.PR↗

arXiv2018

A Novel Co-design Peta-scale Heterogeneous Cluster for Deep Learning Training

Large scale deep Convolution Neural Networks (CNNs) increasingly demands the computing power. It is key for researchers to own a great powerful computing platform to leverage deep learning (DL) advancing.On the other hand, as the commonly-used accelerator, the commodity GPUs cards of new generations are more and more expensive. Consequently, it is of importance to design an affordable distributed heterogeneous system that provides powerful computational capacity and develop a well-suited software that efficiently utilizes its computational capacity. In this paper, we present our co-design distributed system including a peta-scale GPU cluster, called "Manoa". Based on properties and topology of Manoa, we first propose job server framework and implement it, named "MiMatrix". The central node of MiMatrix, referred to as the job server, undertakes all of controlling, scheduling and monitoring, and I/O tasks without weight data transfer for AllReduce processing in each iteration. Therefore, MiMatrix intrinsically solves the bandwidth bottleneck of central node in parameter server framework that is widely used in distributed DL tasks. Meanwhile, we also propose a new AllReduce algorithm, GPUDirect RDMA-Aware AllReduce~(GDRAA), in which both computation and handshake message are O(1) and the number of synchronization is two in each iteration that is a theoretical minimum number. Owe to the dedicated co-design distributed system, MiMatrix efficiently makes use of the Manoa's computational capacity and bandwidth. We benchmark Manoa Resnet50 and Resenet101 on Imagenet-1K dataset. Some of results have demonstrated state-of-the-art.

cs.CV↗

arXiv2018

Large-scale Analysis of Opioid Poisoning Related Hospital Visits in New York State

Opioid related deaths are increasing dramatically in recent years, and opioid epidemic is worsening in the United States. Combating opioid epidemic becomes a high priority for both the U.S. government and local governments such as New York State. Analyzing patient level opioid related hospital visits provides a data driven approach to discover both spatial and temporal patterns and identity potential causes of opioid related deaths, which provides essential knowledge for governments on decision making. In this paper, we analyzed opioid poisoning related hospital visits using New York State SPARCS data, which provides diagnoses of patients in hospital visits. We identified all patients with primary diagnosis as opioid poisoning from 2010-2014 for our main studies, and from 2003-2014 for temporal trend studies. We performed demographical based studies, and summarized the historical trends of opioid poisoning. We used frequent item mining to find co-occurrences of diagnoses for possible causes of poisoning or effects from poisoning. We provided zip code level spatial analysis to detect local spatial clusters, and studied potential correlations between opioid poisoning and demographic and social-economic factors.

q-bio.PE↗

arXiv2018

Generalized Linear Model for Gamma Distributed Variables via Elastic Net Regularization

The Generalized Linear Model (GLM) for the Gamma distribution (glmGamma) is widely used in modeling continuous, non-negative and positive-skewed data, such as insurance claims and survival data. However, model selection for GLM depends on AIC/BIC criteria, which is computationally impractical for even a moderate number of variables. In this paper, we develop variable selection for glmGamma using elastic net regularization (glmGammaNet), for which we provide an algorithm and implementation. The glmGammaNet model is more challening than other more common GLMs as the likelihood function has no global quadratic upper bound, and we develop an efficient accelerated proximal gradient algorithm using a local model. We report simulation study results and discuss the choice of regularization parameter. The method is implemented in the R package glmGammaNet.

stat.ME↗

arXiv2018

On sparsity of the solution to a random quadratic optimization problem

The standard quadratic optimization problem (StQP), i.e. the problem of minimizing a quadratic form $\bold x^TQ\bold x$ on the standard simplex $\{\bold x\ge\bold 0: \bold x^T\bold e=1\}$, is studied. The StQP arises in numerous applications, and it is known to be NP-hard. The first author, Peng and Zhang~\cite{int:Peng-StQP} showed that almost certainly the StQP with a large random matrix $Q=Q^T$, whose upper-triangular entries are i. i. concave-distributed, attains its minimum at a point with few positive components. In this paper we establish sparsity of the solution for a considerably broader class of the distributions, including those supported by $(-\infty,\infty)$, provided that the distribution tail is (super/sub)-exponentially narrow, and also for the matrices $Q=(M+M^T)/2$, when $M$ is not symmetric. {The likely support size in those cases is shown to be polylogarithmic in $n$, the problem dimension.} Following~\cite{int:Peng-StQP} and Chen and Peng ~\cite{ChenPeng2015}, the key ingredients are the first and second order optimality conditions, and the integral bound for the tail distribution of the solution support size. To make these results work for our goal, we obtain a series of estimates involving, in particular, the random interval partitions induced by the order statistics of the elements $Q_{i,j}$.

math.PR↗

arXiv2018

Deep Neural Network Capacity

In recent years, deep neural network exhibits its powerful superiority on information discrimination in many computer vision applications. However, the capacity of deep neural network architecture is still a mystery to the researchers. Intuitively, larger capacity of neural network can always deposit more information to improve the discrimination ability of the model. But, the learnable parameter scale is not feasible to estimate the capacity of deep neural network. Due to the overfitting, directly increasing hidden nodes number and hidden layer number are already demonstrated not necessary to effectively increase the network discrimination ability. In this paper, we propose a novel measurement, named "total valid bits", to evaluate the capacity of deep neural networks for exploring how to quantitatively understand the deep learning and the insights behind its super performance. Specifically, our scheme to retrieve the total valid bits incorporates the skilled techniques in both training phase and inference phase. In the network training, we design decimal weight regularization and 8-bit forward quantization to obtain the integer-oriented network representations. Moreover, we develop adaptive-bitwidth and non-uniform quantization strategy in the inference phase to find the neural network capacity, total valid bits. By allowing zero bitwidth, our adaptive-bitwidth quantization can execute the model reduction and valid bits finding simultaneously. In our extensive experiments, we first demonstrate that our total valid bits is a good indicator of neural network capacity. We also analyze the impact on network capacity from the network architecture and advanced training skills, such as dropout and batch normalization.

cs.CV↗

arXiv2018

Fully-Coupled Two-Stream Spatiotemporal Networks for Extremely Low Resolution Action Recognition

A major emerging challenge is how to protect people's privacy as cameras and computer vision are increasingly integrated into our daily lives, including in smart devices inside homes. A potential solution is to capture and record just the minimum amount of information needed to perform a task of interest. In this paper, we propose a fully-coupled two-stream spatiotemporal architecture for reliable human action recognition on extremely low resolution (e.g., 12x16 pixel) videos. We provide an efficient method to extract spatial and temporal features and to aggregate them into a robust feature representation for an entire action video sequence. We also consider how to incorporate high resolution videos during training in order to build better low resolution action recognition models. We evaluate on two publicly-available datasets, showing significant improvements over the state-of-the-art.

cs.CV↗

arXiv2018

High Dimensional Elliptical Sliced Inverse Regression in non-Gaussian Distributions

Sliced inverse regression (SIR) is the most widely-used sufficient dimension reduction method due to its simplicity, generality and computational efficiency. However, when the distribution of the covariates deviates from the multivariate normal distribution, the estimation efficiency of SIR is rather low. In this paper, we propose a robust alternative to SIR - called elliptical sliced inverse regression (ESIR) for analysing high dimensional, elliptically distributed data. There are wide range of applications of the elliptically distributed data, especially in finance and economics where the distribution of the data is often heavy-tailed. To tackle the heavy-tailed elliptically distributed covariates, we novelly utilize the multivariate Kendall's tau matrix in a framework of so-called generalized eigenvector problem for sufficient dimension reduction. Methodologically, we present a practical algorithm for our method. Theoretically, we investigate the asymptotic behavior of the ESIR estimator and obtain the corresponding convergence rate under high dimensional setting. Quantities of simulation results show that ESIR significantly improves the estimation efficiency in heavy-tailed scenarios. A stock exchange data analysis also demonstrates the effectiveness of our method. Moreover, ESIR can be easily extended to most other sufficient dimension reduction methods.

stat.ME↗

arXiv2017

Distribution Regression

Linear regression is a fundamental and popular statistical method. There are various kinds of linear regression, such as mean regression and quantile regression. In this paper, we propose a new one called distribution regression, which allows broad-spectrum of the error distribution in the linear regression. Our method uses nonparametric technique to estimate regression parameters. Our studies indicate that our method provides a better alternative than mean regression and quantile regression under many settings, particularly for asymmetrical heavy-tailed distribution or multimodal distribution of the error term. Under some regular conditions, our estimator is $\sqrt n$-consistent and possesses the asymptotically normal distribution. The proof of the asymptotic normality of our estimator is very challenging because our nonparametric likelihood function cannot be transformed into sum of independent and identically distributed random variables. Furthermore, penalized likelihood estimator is proposed and enjoys the so-called oracle property with diverging number of parameters. Numerical studies also demonstrate the effectiveness and the flexibility of the proposed method.

stat.ME↗

arXiv2017

ArbiText: Arbitrary-Oriented Text Detection in Unconstrained Scene

Arbitrary-oriented text detection in the wild is a very challenging task, due to the aspect ratio, scale, orientation, and illumination variations. In this paper, we propose a novel method, namely Arbitrary-oriented Text (or ArbText for short) detector, for efficient text detection in unconstrained natural scene images. Specifically, we first adopt the circle anchors rather than the rectangular ones to represent bounding boxes, which is more robust to orientation variations. Subsequently, we incorporate a pyramid pooling module into the Single Shot MultiBox Detector framework, in order to simultaneously explore the local and global visual information, which can, therefore, generate more confidential detection results. Experiments on established scene-text datasets, such as the ICDAR 2015 and MSRA-TD500 datasets, have demonstrated the supe rior performance of the proposed method, compared to the state-of-the-art approaches.

cs.CV↗

arXiv2017

Sparse Photometric 3D Face Reconstruction Guided by Morphable Models

We present a novel 3D face reconstruction technique that leverages sparse photometric stereo (PS) and latest advances on face registration/modeling from a single image. We observe that 3D morphable faces approach provides a reasonable geometry proxy for light position calibration. Specifically, we develop a robust optimization technique that can calibrate per-pixel lighting direction and illumination at a very high precision without assuming uniform surface albedos. Next, we apply semantic segmentation on input images and the geometry proxy to refine hairy vs. bare skin regions using tailored filters. Experiments on synthetic and real data show that by using a very small set of images, our technique is able to reconstruct fine geometric details such as wrinkles, eyebrows, whelks, pores, etc, comparable to and sometimes surpassing movie quality productions.

cs.CV↗

arXiv2017

Search for CP violation effects in the $h\to ττ$ decay with $e^+e^-$ colliders

A new method is proposed to reconstruct the neutrinos in the $e^+e^-\to Zh$ process followed by the $h\toττ$ decay. With the help of a refined Higgs momentum reconstruction from the recoiling system and the impact parameters, high precision in the determination of the momentum of neutrinos can be achieved. The prospect of measuring the Higgs CP mixing angle with the $h\toττ$ decay at future $e^+e^-$ colliders is studied with the new method. The analysis is based on a detailed detector simulation of the signal and backgrounds. The fully reconstructed neutrinos and also other visible products from the tau decay are used to build matrix element (ME) based CP observables. With 5 $ab^{-1}$ of data at $E_{\text{CM}}=250$ GeV, a precision of $2.9^\circ$ can be achieved for the CP mixing angle with three main one-prong decay modes of the taus. The precision is found to be about 35\% better than the other methods.

hep-ph↗