arXiv subjects

Chao Gao

Publications and source records attributed to Chao Gao.

At least 145 records · Page 8Linked to original sources

arXiv2019

Skynet: A Top Deep RL Agent in the Inaugural Pommerman Team Competition

The Pommerman Team Environment is a recently proposed benchmark which involves a multi-agent domain with challenges such as partial observability, decentralized execution (without communication), and very sparse and delayed rewards. The inaugural Pommerman Team Competition held at NeurIPS 2018 hosted 25 participants who submitted a team of 2 agents. Our submission nn_team_skynet955_skynet955 won 2nd place of the "learning agents'' category. Our team is composed of 2 neural networks trained with state of the art deep reinforcement learning algorithms and makes use of concepts like reward shaping, curriculum learning, and an automatic reasoning module for action pruning. Here, we describe these elements and additionally we present a collection of open-sourced agents that can be used for training and testing in the Pommerman environment. Code available at: https://github.com/BorealisAI/pommerman-baseline

cs.MA↗

arXiv2019

Safer Deep RL with Shallow MCTS: A Case Study in Pommerman

Safe reinforcement learning has many variants and it is still an open research problem. Here, we focus on how to use action guidance by means of a non-expert demonstrator to avoid catastrophic events in a domain with sparse, delayed, and deceptive rewards: the recently-proposed multi-agent benchmark of Pommerman. This domain is very challenging for reinforcement learning (RL) --- past work has shown that model-free RL algorithms fail to achieve significant learning. In this paper, we shed light into the reasons behind this failure by exemplifying and analyzing the high rate of catastrophic events (i.e., suicides) that happen under random exploration in this domain. While model-free random exploration is typically futile, we propose a new framework where even a non-expert simulated demonstrator, e.g., planning algorithms such as Monte Carlo tree search with small number of rollouts, can be integrated to asynchronous distributed deep reinforcement learning methods. Compared to vanilla deep RL algorithms, our proposed methods both learn faster and converge to better policies on a two-player mini version of the Pommerman game.

cs.LG↗

arXiv2019

Generative Adversarial Nets for Robust Scatter Estimation: A Proper Scoring Rule Perspective

Robust scatter estimation is a fundamental task in statistics. The recent discovery on the connection between robust estimation and generative adversarial nets (GANs) by Gao et al. (2018) suggests that it is possible to compute depth-like robust estimators using similar techniques that optimize GANs. In this paper, we introduce a general learning via classification framework based on the notion of proper scoring rules. This framework allows us to understand both matrix depth function and various GANs through the lens of variational approximations of $f$-divergences induced by proper scoring rules. We then propose a new class of robust scatter estimators in this framework by carefully constructing discriminators with appropriate neural network structures. These estimators are proved to achieve the minimax rate of scatter estimation under Huber's contamination model. Our numerical results demonstrate its good performance under various settings against competitors in the literature.

cs.LG↗

arXiv2019

Robust Estimation and Generative Adversarial Nets

Robust estimation under Huber's $ε$-contamination model has become an important topic in statistics and theoretical computer science. Statistically optimal procedures such as Tukey's median and other estimators based on depth functions are impractical because of their computational intractability. In this paper, we establish an intriguing connection between $f$-GANs and various depth functions through the lens of $f$-Learning. Similar to the derivation of $f$-GANs, we show that these depth functions that lead to statistically optimal robust estimators can all be viewed as variational lower bounds of the total variation distance in the framework of $f$-Learning. This connection opens the door of computing robust estimators using tools developed for training GANs. In particular, we show in both theory and experiments that some appropriate structures of discriminator networks with hidden layers in GANs lead to statistically optimal robust location estimators for both Gaussian distribution and general elliptical distributions where first moment may not exist.

stat.ML↗

arXiv2019

Minimax Rates in Network Analysis: Graphon Estimation, Community Detection and Hypothesis Testing

This paper surveys some recent developments in fundamental limits and optimal algorithms for network analysis. We focus on minimax optimal rates in three fundamental problems of network analysis: graphon estimation, community detection, and hypothesis testing. For each problem, we review state-of-the-art results in the literature followed by general principles behind the optimal procedures that lead to minimax estimation and testing. This allows us to connect problems in network analysis to other statistical inference problems from a general perspective.

math.ST↗

arXiv2019

Dynamical Fractal in Quantum Gases with Discrete Scaling Symmetry

Inspired by the similarity between the fractal Weierstrass function and quantum systems with discrete scaling symmetry, we establish general conditions under which the dynamics of a quantum system will exhibit fractal structure in the time domain. As an example, we discuss the dynamics of the Loschmidt amplitude and the zero-momentum occupation of a single particle moving in a scale invariant $1/r^2$ potential. In order to show these conditions can be realized in ultracold atomic gases, we perform numerical simulations with practical experimental parameters, which shows that the dynamical fractal can be observed in realistic time scales. The predication can be directly verified in current cold atom experiments.

cond-mat.quant-gas↗

arXiv2018

An Attention-Gated Convolutional Neural Network for Sentence Classification

The classification of sentences is very challenging, since sentences contain the limited contextual information. In this paper, we proposed an Attention-Gated Convolutional Neural Network (AGCNN) for sentence classification, which generates attention weights from the feature's context windows of different sizes by using specialized convolution encoders. It makes full use of limited contextual information to extract and enhance the influence of important features in predicting the sentence's category. Experimental results demonstrated that our model can achieve up to 3.1% higher accuracy than standard CNN models, and gain competitive results over the baselines on four out of the six tasks. Besides, we designed an activation function, namely, Natural Logarithm rescaled Rectified Linear Unit (NLReLU). Experiments showed that NLReLU can outperform ReLU and is comparable to other well-known activation functions on AGCNN.

cs.CL↗

arXiv2018

Continual Match Based Training in Pommerman: Technical Report

Continual learning is the ability of agents to improve their capacities throughout multiple tasks continually. While recent works in the literature of continual learning mostly focused on developing either particular loss functions or specialized structures of neural network explaining the episodic memory or neural plasticity, we study continual learning from the perspective of the training mechanism. Specifically, we propose a COnitnual Match BAsed Training (COMBAT) framework for training a population of advantage-actor-critic (A2C) agents in Pommerman, a partially observable multi-agent environment with no communication. Following the COMBAT framework, we trained an agent, namely, Navocado, that won the title of the top 1 learning agent in the NeurIPS 2018 Pommerman Competition. Two critical features of our agent are worth mentioning. Firstly, our agent did not learn from any demonstrations. Secondly, our agent is highly reproducible. As a technical report, we articulate the design of state space, action space, reward, and most importantly, the COMBAT framework for our Pommerman agent. We show in the experiments that Pommerman is a perfect environment for studying continual learning, and the agent can improve its performance by continually learning new skills without forgetting the old ones. Finally, the result in the Pommerman Competition verifies the robustness of our agent when competing with various opponents.

cs.AI↗

arXiv2018

Steady-state phase diagram of quantum gases in a lattice coupled to a membrane

In a recent experiment [Vochezer {\it et al.,} Phys. Rev. Lett. \textbf{120}, 073602 (2018)], a novel kind of hybrid atom-opto-mechanical system has been realized by coupling atoms in a lattice to a membrane. While such system promises a viable contender in the competitive field of simulating non-equilibrium many-body physics, its complete steady-state phase diagram is still lacking. Here we study the phase diagram of this hybrid system based on an atomic Bose-Hubbard model coupled to a quantum harmonic oscillator. We take both the expectation value of the bosonic operator and the mechanical motion of the membrane as order parameters, and thereby identify four different quantum phases. Importantly, we find the atomic gas in the steady state of such non-equilibrium setting undergoes a superfluid-Mott-insulator transition when the atom-membrane coupling is tuned to increase. Such steady-state phase transition can be seen as the non-equilibrium counterpart of the conventional superfluid-Mott-insulator transition in the ground state of Bose-Hubbard model. Further, no matter which phase the quantum gas is in, the mechanic motion of the membrane exhibits a transition from an incoherent vibration to a coherent one when the atom-membrane coupling increases, agreeing with the experimental observations. Our present study provides a simple way to study non-equilibrium many-body physics that is complementary to ongoing experiments on the hybrid atomic and opto-mechanical systems.

cond-mat.quant-gas↗

arXiv2018

Mixing Time of Metropolis-Hastings for Bayesian Community Detection

We study the computational complexity of a Metropolis-Hastings algorithm for Bayesian community detection. We first establish a posterior strong consistency result for a natural prior distribution on stochastic block models under the optimal signal-to-noise ratio condition in the literature. We then give a set of conditions that guarantee rapid mixing of a simple Metropolis-Hastings algorithm. The mixing time analysis is based on a careful study of posterior ratios and a canonical path argument to control the spectral gap of the Markov chain.

math.ST↗

arXiv2018

Multi-resolution Progressive Computational Ghost Imaging

Ghost imaging needs massive measurements to obtain an image with good visibility and the imaging speed is usually very low. In order to realize real-time high-resolution ghost imaging of a target which is located in a scenario with a large field of view (FOV), we propose a high-speed multi-resolution progressive computational ghost imaging approach. The target area is firstly locked by a low-resolution image with a small number of measurements, then high-resolution imaging of the target can be obtained by only modulating the light fields corresponding to the target area. The experiments verify the feasibility of the approach. The influence of detection signal-to-noise ratio on the quality of multi-resolution progressive computational ghost imaging is also investigated experimentally. This approach may be applied to some practical application scenarios such as ground-to-air or air-to-air imaging with a large FOV.

eess.IV↗

arXiv2018

Optimal Estimation and Completion of Matrices with Biclustering Structures

Biclustering structures in data matrices were first formalized in a seminal paper by John Hartigan (1972) where one seeks to cluster cases and variables simultaneously. Such structures are also prevalent in block modeling of networks. In this paper, we develop a unified theory for the estimation and completion of matrices with biclustering structures, where the data is a partially observed and noise contaminated data matrix with a certain biclustering structure. In particular, we show that a constrained least squares estimator achieves minimax rate-optimal performance in several of the most important scenarios. To this end, we derive unified high probability upper bounds for all sub-Gaussian data and also provide matching minimax lower bounds in both Gaussian and binary cases. Due to the close connection of graphon to stochastic block models, an immediate consequence of our general results is a minimax rate-optimal estimator for sparse graphons.

math.ST↗

arXiv2018

Urban contact structures for epidemic simulations: Correcting biases in data-driven approaches

Epidemics are emergent phenomena depending on the epidemiological characteristics of pathogens and the interaction and movement of people. Public transit systems have provided much important information about the movement of people, but there are also other means of transportation (e.g., bicycle and private car), that are invisible to public transit data. This discrepancy can induce a bias in disease models that leads to mispredictions of epidemic growth (e.g., peak prevalence and time). In our study, we aim to advance and compare the epidemic spreading dynamics using public transit trips, in contrast to more accurate estimates of population movement using mobile phones traces. In our study, we simulate epidemic outbreaks in a cohort of two million mobile phone users. We use a metapopulation model incorporating susceptible-infected-recovered dynamics to analyze and compare different effective contract matrices, constructed by the public transit systems and mobile phones respectively, on the process of epidemics. We find that epidemic outbreaks using public transit trips tend to be underestimated in terms of the epidemic spreading dynamics, reaching epidemic peaks weaker and later. This is rooted in a later introduction of new infectious people into uninfected locations.

cs.SI↗

arXiv2018

Ghost imaging for an occluded object

Imaging for an occluded object is usually a difficult problem, in this letter, we introduce an imaging scheme based on computational ghost imaging, which can obtain the image of a target object behind an obstacle. According to our theoretical analysis, once the distance between the object and the obstacle is far enough, one can obtain the image of the object by using ghost imaging technique. The wavelength of the light source also affects the quality of the reconstructed image. In addition, if the bucket detector is placed far away from the obstacle, a tiny point-like detector without collecting lens can be applied to realize the imaging. These theoretical results above have been verified with our numerical simulations. Furthermore, the robustness of this imaging scheme is also investigated.

physics.optics↗

arXiv2018

Fulde-Ferrell-Larkin-Ovchinnikov pairing states of a polarized dipolar Fermi gas trapped in a one-dimensional optical lattice

We study the interplay between the long- and short-range interaction of a one-dimensional optical lattice system of two-component dipolar fermions by using the density matrix renormalization group method. The atomic density profile, pairing-pairing correlation function, and the compressibility are calculated in the ground state, from which we identify the parameter region of the Fulde-Ferrell-Larkin-Ovchinnikov (FFLO) pairing state, half-metal (HM) state, FFLO-HM state, and the normal polarized state, and thus the phase diagram in the coordinates of the long- and short-range interaction strength. The effect of the long-range dipolar interaction on the FFLO state is discussed in details. We find that the long-range part of the dipole-dipole interaction does not sweep away the FFLO superconducting region that is driven by the short-range interaction in the Hubbard model, and thus the FFLO state survives in the wide parameter space of the long-range interaction, polarization and the filling.

cond-mat.quant-gas↗

arXiv2018

Critical behavior of order parameter at the nonequilibrium phase transition of the Ising model

After a quench of transverse field, the asymptotic long-time state of Ising model displays a transition from a ferromagnetic phase to a paramagnetic phase as the post-quench field strength increases, which is revealed by the vanishing of the order parameter defined as the averaged magnetization over time. We estimate the critical behavior of the magnetization at this nonequilibrium phase transition by using mean-field approximation. In the vicinity of the critical field, the magnetization vanishes as the inverse of a logarithmic function, which is significantly distinguished from the critical behavior of order parameter at the corresponding equilibrium phase transition, i.e. a power-law function.

cond-mat.quant-gas↗

arXiv2018

A General Framework for Bayes Structured Linear Models

High dimensional statistics deals with the challenge of extracting structured information from complex model settings. Compared with the growing number of frequentist methodologies, there are rather few theoretically optimal Bayes methods that can deal with very general high dimensional models. In contrast, Bayes methods have been extensively studied in various nonparametric settings and rate optimal posterior contraction results have been established. This paper provides a unified approach to both Bayes high dimensional statistics and Bayes nonparametrics in a general framework of structured linear models. With the proposed two-step model selection prior, we prove a general theorem of posterior contraction under an abstract setting. The main theorem can be used to derive new results on optimal posterior contraction under many complex model settings including stochastic block model, graphon estimation and dictionary learning. It can also be used to re-derive optimal posterior contraction for problems such as sparse linear regression and nonparametric aggregation, which improve upon previous Bayes results for these problems. The key of the success lies in the proposed two-step prior distribution. The prior on the parameters is an elliptical Laplace distribution that is capable to model signals with large magnitude, and the prior on the models involves an important correction factor that compensates the effect of the normalizing constant of the elliptical Laplace distribution.

math.ST↗

arXiv2018

Goodness-of-Fit Tests for Random Partitions via Symmetric Polynomials

We consider goodness-of-fit tests with i.i.d. samples generated from a categorical distribution $(p_1,...,p_k)$. For a given $(q_1,...,q_k)$, we test the null hypothesis whether $p_j=q_{π(j)}$ for some label permutation $π$. The uncertainty of label permutation implies that the null hypothesis is composite instead of being singular. In this paper, we construct a testing procedure using statistics that are defined as indefinite integrals of some symmetric polynomials. This method is aimed directly at the invariance of the problem, and avoids the need of matching the unknown labels. The asymptotic distribution of the testing statistic is shown to be chi-squared, and its power is proved to be nearly optimal under a local alternative hypothesis. Various degenerate structures of the null hypothesis are carefully analyzed in the paper. A two-sample version of the test is also studied.

math.ST↗