arXiv subjects

Xiang Zhou

Publications and source records attributed to Xiang Zhou.

At least 127 records · Page 7Linked to original sources

arXiv2021

Value-Gradient based Formulation of Optimal Control Problem and Machine Learning Algorithm

Optimal control problem is typically solved by first finding the value function through Hamilton-Jacobi equation (HJE) and then taking the minimizer of the Hamiltonian to obtain the control. In this work, instead of focusing on the value function, we propose a new formulation for the gradient of the value function (value-gradient) as a decoupled system of partial differential equations in the context of continuous-time deterministic discounted optimal control problem. We develop an efficient iterative scheme for this system of equations in parallel by utilizing the properties that they share the same characteristic curves as the HJE for the value function. For the theoretical part, we prove that this iterative scheme converges linearly in $L_α^2$ sense for some suitable exponent $α$ in a weight function. For the numerical method, we combine characteristic line method with machine learning techniques. Specifically, we generate multiple characteristic curves at each policy iteration from an ensemble of initial states, and compute both the value function and its gradient simultaneously on each curve as the labelled data. Then supervised machine learning is applied to minimize the weighted squared loss for both the value function and its gradients. Experimental results demonstrate that this new method not only significantly increases the accuracy but also improves the efficiency and robustness of the numerical estimates, particularly with less amount of characteristics data or fewer training steps.

math.OC↗

arXiv2021

Composing Answer from Multi-spans for Reading Comprehension

This paper presents a novel method to generate answers for non-extraction machine reading comprehension (MRC) tasks whose answers cannot be simply extracted as one span from the given passages. Using a pointer network-style extractive decoder for such type of MRC may result in unsatisfactory performance when the ground-truth answers are given by human annotators or highly re-paraphrased from parts of the passages. On the other hand, using generative decoder cannot well guarantee the resulted answers with well-formed syntax and semantics when encountering long sentences. Therefore, to alleviate the obvious drawbacks of both sides, we propose an answer making-up method from extracted multi-spans that are learned by our model as highly confident $n$-gram candidates in the given passage. That is, the returned answers are composed of discontinuous multi-spans but not just one consecutive span in the given passages anymore. The proposed method is simple but effective: empirical experiments on MS MARCO show that the proposed method has a better performance on accurately generating long answers, and substantially outperforms two competitive typical one-span and Seq2Seq baseline decoders.

cs.CL↗

arXiv2021

Verifying Verified Code

A recent case study from AWS by Chong et al. proposes an effective methodology for Bounded Model Checking in industry. In this paper, we report on a follow up case study that explores the methodology from the perspective of three research questions: (a) can proof artifacts be used across verification tools; (b) are there bugs in verified code; and (c) can specifications be improved. To study these questions, we port the verification tasks for $\texttt{aws-c-common}$ library to SEAHORN and KLEE. We show the benefits of using compiler semantics and cross-checking specifications with different verification techniques, and call for standardizing proof library extensions to increase specification reuse. The verification tasks discussed are publicly available online.

cs.SE↗

arXiv2021

Spin Flavor Spectral Splits of Supernova Neutrino Flavor Conversions

The supernova neutrino flavor evolution in the presence of the non-trivial neutrino magnetic moment and strong magnetic field is numerically derived using the two-flavor and single-angle approximation. The novel properties of collective neutrino oscillations are studied and distinct patterns of flavor and spin-flavor spectral splits are presented. Finally we also discuss how the neutrino magnetic moment affects the observable supernova neutrino energy spectra.

hep-ph↗

arXiv2021

Hidden Biases in Unreliable News Detection Datasets

Automatic unreliable news detection is a research problem with great potential impact. Recently, several papers have shown promising results on large-scale news datasets with models that only use the article itself without resorting to any fact-checking mechanism or retrieving any supporting evidence. In this work, we take a closer look at these datasets. While they all provide valuable resources for future research, we observe a number of problems that may lead to results that do not generalize in more realistic settings. Specifically, we show that selection bias during data collection leads to undesired artifacts in the datasets. In addition, while most systems train and predict at the level of individual articles, overlapping article sources in the training and evaluation data can provide a strong confounding factor that models can exploit. In the presence of this confounding factor, the models can achieve good performance by directly memorizing the site-label mapping instead of modeling the real task of unreliable news detection. We observed a significant drop (>10%) in accuracy for all models tested in a clean split with no train/test source overlap. Using the observations and experimental results, we provide practical suggestions on how to create more reliable datasets for the unreliable news detection task. We suggest future dataset creation include a simple model as a difficulty/bias probe and future model development use a clean non-overlapping site and date split.

cs.CL↗

arXiv2021

Electron-Ion Collider in China

Lepton scattering is an established ideal tool for studying inner structure of small particles such as nucleons as well as nuclei. As a future high energy nuclear physics project, an Electron-ion collider in China (EicC) has been proposed. It will be constructed based on an upgraded heavy-ion accelerator, High Intensity heavy-ion Accelerator Facility (HIAF) which is currently under construction, together with a new electron ring. The proposed collider will provide highly polarized electrons (with a polarization of $\sim$80%) and protons (with a polarization of $\sim$70%) with variable center of mass energies from 15 to 20 GeV and the luminosity of (2-3) $\times$ 10$^{33}$ cm$^{-2}$ s$^{-1}$. Polarized deuterons and Helium-3, as well as unpolarized ion beams from Carbon to Uranium, will be also available at the EicC. The main foci of the EicC will be precision measurements of the structure of the nucleon in the sea quark region, including 3D tomography of nucleon; the partonic structure of nuclei and the parton interaction with the nuclear environment; the exotic states, especially those with heavy flavor quark contents. In addition, issues fundamental to understanding the origin of mass could be addressed by measurements of heavy quarkonia near-threshold production at the EicC. In order to achieve the above-mentioned physics goals, a hermetical detector system will be constructed with cutting-edge technologies. This document is the result of collective contributions and valuable inputs from experts across the globe. The EicC physics program complements the ongoing scientific programs at the Jefferson Laboratory and the future EIC project in the United States. The success of this project will also advance both nuclear and particle physics as well as accelerator and detector technology in China.

nucl-ex↗

arXiv2021

Regime Switching Bandits

We study a multi-armed bandit problem where the rewards exhibit regime switching. Specifically, the distributions of the random rewards generated from all arms are modulated by a common underlying state modeled as a finite-state Markov chain. The agent does not observe the underlying state and has to learn the transition matrix and the reward distributions. We propose a learning algorithm for this problem, building on spectral method-of-moments estimations for hidden Markov models, belief error control in partially observable Markov decision processes and upper-confidence-bound methods for online learning. We also establish an upper bound $O(T^{2/3}\sqrt{\log T})$ for the proposed learning algorithm where $T$ is the learning horizon. Finally, we conduct proof-of-concept experiments to illustrate the performance of the learning algorithm.

cs.LG↗

arXiv2020

Fine-grained Attention and Feature-sharing Generative Adversarial Networks for Single Image Super-Resolution

The traditional super-resolution methods that aim to minimize the mean square error usually produce the images with over-smoothed and blurry edges, due to the lose of high-frequency details. In this paper, we propose two novel techniques in the generative adversarial networks to produce photo-realistic images for image super-resolution. Firstly, instead of producing a single score to discriminate images between real and fake, we propose a variant, called Fine-grained Attention Generative Adversarial Network for image super-resolution (FASRGAN), to discriminate each pixel between real and fake. FASRGAN adopts a Unet-like network as the discriminator with two outputs: an image score and an image score map. The score map has the same spatial size as the HR/SR images, serving as the fine-grained attention to represent the degree of reconstruction difficulty for each pixel. Secondly, instead of using different networks for the generator and the discriminator in the SR problem, we use a feature-sharing network (Fs-SRGAN) for both the generator and the discriminator. By network sharing, certain information is shared between the generator and the discriminator, which in turn can improve the ability of producing high-quality images. Quantitative and visual comparisons with the state-of-the-art methods on the benchmark datasets demonstrate the superiority of our methods. The application of super-resolution images to object recognition further proves that the proposed methods endow the power to reconstruction capabilities and the excellent super-resolution effects.

eess.IV↗

arXiv2020

Filling the Gap of Utterance-aware and Speaker-aware Representation for Multi-turn Dialogue

A multi-turn dialogue is composed of multiple utterances from two or more different speaker roles. Thus utterance- and speaker-aware clues are supposed to be well captured in models. However, in the existing retrieval-based multi-turn dialogue modeling, the pre-trained language models (PrLMs) as encoder represent the dialogues coarsely by taking the pairwise dialogue history and candidate response as a whole, the hierarchical information on either utterance interrelation or speaker roles coupled in such representations is not well addressed. In this work, we propose a novel model to fill such a gap by modeling the effective utterance-aware and speaker-aware representations entailed in a dialogue history. In detail, we decouple the contextualized word representations by masking mechanisms in Transformer-based PrLM, making each word only focus on the words in current utterance, other utterances, two speaker roles (i.e., utterances of sender and utterances of receiver), respectively. Experimental results show that our method boosts the strong ELECTRA baseline substantially in four public benchmark datasets, and achieves various new state-of-the-art performance over previous methods. A series of ablation studies are conducted to demonstrate the effectiveness of our method.

cs.CL↗

arXiv2020

Some Doubly and Multiply Robust Estimators of Controlled Direct Effects

This letter introduces several doubly, triply, and quadruply robust estimators of the controlled direct effect. Among them, the triply and quadruply robust estimators are locally semiparametric efficient, and well suited to the use of data-adaptive methods for estimating their nuisance functions.

stat.ME↗

arXiv2020

The Curse of Performance Instability in Analysis Datasets: Consequences, Source, and Suggestions

We find that the performance of state-of-the-art models on Natural Language Inference (NLI) and Reading Comprehension (RC) analysis/stress sets can be highly unstable. This raises three questions: (1) How will the instability affect the reliability of the conclusions drawn based on these analysis sets? (2) Where does this instability come from? (3) How should we handle this instability and what are some potential solutions? For the first question, we conduct a thorough empirical study over analysis sets and find that in addition to the unstable final performance, the instability exists all along the training curve. We also observe lower-than-expected correlations between the analysis validation set and standard validation set, questioning the effectiveness of the current model-selection routine. Next, to answer the second question, we give both theoretical explanations and empirical evidence regarding the source of the instability, demonstrating that the instability mainly comes from high inter-example correlations within analysis sets. Finally, for the third question, we discuss an initial attempt to mitigate the instability and suggest guidelines for future work such as reporting the decomposed variance for more interpretable results and fair comparison across models. Our code is publicly available at: https://github.com/owenzx/InstabilityAnalysis

cs.CL↗

arXiv2020

Projection Method for Saddle Points of Energy Functional in $H^{-1}$ Metric

Saddle points play important roles as the transition states of activated process in gradient system driven by energy functional. However, for the same energy functional, the saddle points, as well as other stationary points, are different in different metrics such as the $L^2$ metric and the $H^{-1}$ metric. The saddle point calculation in $H^{-1}$ metric is more challenging with much higher computational cost since it involves higher order derivative in space and the inner product calculation needs to solve another Possion equation to get the $Δ^{-1}$ operator. In this paper, we introduce the projection idea to the existing saddle point search methods, gentlest ascent dynamics (GAD) and iterative minimization formulation (IMF), to overcome this numerical challenge due to $H^{-1}$ metric. Our new method in the $L^2$ metric only by carefully incorporates a simple linear projection step. We show that our projection method maintains the same convergence speed of the original GAD and IMF, but the new algorithm is much faster than the direct method for $H^{-1}$ problem. The numerical results of saddle points in the one dimensional Ginzburg-Landau free energy and the two dimensional Landau-Brazovskii free energy in $H^{-1}$ metric are presented to demonstrate the efficiency of this new method.

math.NA↗

arXiv2020

Explicit Estimation of Derivatives from Data and Differential Equations by Gaussian Process Regression

In this work, we employ the Bayesian inference framework to solve the problem of estimating the solution and particularly, its derivatives, which satisfy a known differential equation, from the given noisy and scarce observations of the solution data only. To address the key issue of accuracy and robustness of derivative estimation, we use the Gaussian processes to jointly model the solution, the derivatives, and the differential equation. By regarding the linear differential equation as a linear constraint, a Gaussian process regression with constraint method (GPRC) is developed to improve the accuracy of prediction of derivatives. For nonlinear differential equations, we propose a Picard-iteration-like approximation of linearization around the Gaussian process obtained only from data so that our GPRC can be still iteratively applicable. Besides, a product of experts method is applied to ensure the initial or boundary condition is considered to further enhance the prediction accuracy of the derivatives. We present several numerical results to illustrate the advantages of our new method in comparison to the standard data-driven Gaussian process regression.

stat.CO↗

arXiv2020

What Can We Learn from Collective Human Opinions on Natural Language Inference Data?

Despite the subjective nature of many NLP tasks, most NLU evaluations have focused on using the majority label with presumably high agreement as the ground truth. Less attention has been paid to the distribution of human opinions. We collect ChaosNLI, a dataset with a total of 464,500 annotations to study Collective HumAn OpinionS in oft-used NLI evaluation sets. This dataset is created by collecting 100 annotations per example for 3,113 examples in SNLI and MNLI and 1,532 examples in Abductive-NLI. Analysis reveals that: (1) high human disagreement exists in a noticeable amount of examples in these datasets; (2) the state-of-the-art models lack the ability to recover the distribution over human labels; (3) models achieve near-perfect accuracy on the subset of data with a high level of human agreement, whereas they can barely beat a random guess on the data with low levels of human agreement, which compose most of the common errors made by state-of-the-art models on the evaluation sets. This questions the validity of improving model performance on old metrics for the low-agreement part of evaluation datasets. Hence, we argue for a detailed examination of human agreement in future data collection efforts, and evaluating model outputs against the distribution over collective human opinions. The ChaosNLI dataset and experimental scripts are available at https://github.com/easonnie/ChaosNLI

cs.CL↗

arXiv2020

Bayesian Hierarchical Models for High-Dimensional Mediation Analysis with Coordinated Selection of Correlated Mediators

We consider Bayesian high-dimensional mediation analysis to identify among a large set of correlated potential mediators the active ones that mediate the effect from an exposure variable to an outcome of interest. Correlations among mediators are commonly observed in modern data analysis; examples include the activated voxels within connected regions in brain image data, regulatory signals driven by gene networks in genome data and correlated exposure data from the same source. When correlations are present among active mediators, mediation analysis that fails to account for such correlation can be sub-optimal and may lead to a loss of power in identifying active mediators. Building upon a recent high-dimensional mediation analysis framework, we propose two Bayesian hierarchical models, one with a Gaussian mixture prior that enables correlated mediator selection and the other with a Potts mixture prior that accounts for the correlation among active mediators in mediation analysis. We develop efficient sampling algorithms for both methods. Various simulations demonstrate that our methods enable effective identification of correlated active mediators, which could be missed by using existing methods that assume prior independence among active mediators. The proposed methods are applied to the LIFECODES birth cohort and the Multi-Ethnic Study of Atherosclerosis (MESA) and identified new active mediators with important biological implications.

stat.AP↗

arXiv2020

Deep Reinforcement Learning for On-line Dialogue State Tracking

Dialogue state tracking (DST) is a crucial module in dialogue management. It is usually cast as a supervised training problem, which is not convenient for on-line optimization. In this paper, a novel companion teaching based deep reinforcement learning (DRL) framework for on-line DST optimization is proposed. To the best of our knowledge, this is the first effort to optimize the DST module within DRL framework for on-line task-oriented spoken dialogue systems. In addition, dialogue policy can be further jointly updated. Experiments show that on-line DST optimization can effectively improve the dialogue manager performance while keeping the flexibility of using predefined policy. Joint training of both DST and policy can further improve the performance.

cs.CL↗

arXiv2020

Bayesian Sparse Mediation Analysis with Targeted Penalization of Natural Indirect Effects

Causal mediation analysis aims to characterize an exposure's effect on an outcome and quantify the indirect effect that acts through a given mediator or a group of mediators of interest. With the increasing availability of measurements on a large number of potential mediators, like the epigenome or the microbiome, new statistical methods are needed to simultaneously accommodate high-dimensional mediators while directly target penalization of the natural indirect effect (NIE) for active mediator identification. Here, we develop two novel prior models for identification of active mediators in high-dimensional mediation analysis through penalizing NIEs in a Bayesian paradigm. Both methods specify a joint prior distribution on the exposure-mediator effect and mediator-outcome effect with either (a) a four-component Gaussian mixture prior or (b) a product threshold Gaussian prior. By jointly modeling the two parameters that contribute to the NIE, the proposed methods enable penalization on their product in a targeted way. Resultant inference can take into account the four-component composite structure underlying the NIE. We show through simulations that the proposed methods improve both selection and estimation accuracy compared to other competing methods. We applied our methods for an in-depth analysis of two ongoing epidemiologic studies: the Multi-Ethnic Study of Atherosclerosis (MESA) and the LIFECODES birth cohort. The identified active mediators in both studies reveal important biological pathways for understanding disease mechanisms.

stat.AP↗

arXiv2020

Machine Learning and Control Theory

We survey in this article the connections between Machine Learning and Control Theory. Control Theory provide useful concepts and tools for Machine Learning. Conversely Machine Learning can be used to solve large control problems. In the first part of the paper, we develop the connections between reinforcement learning and Markov Decision Processes, which are discrete time control problems. In the second part, we review the concept of supervised learning and the relation with static optimization. Deep learning which extends supervised learning, can be viewed as a control problem. In the third part, we present the links between stochastic gradient descent and mean-field theory. Conversely, in the fourth and fifth parts, we review machine learning approaches to stochastic control problems, and focus on the deterministic case, to explain, more easily, the numerical algorithms.

cs.LG↗