arXiv subjects

Ya-Ping Hsieh

Publications and source records attributed to Ya-Ping Hsieh.

29 records · Page 2Linked to original sources

arXiv2020

Mirrored Langevin Dynamics

We consider the problem of sampling from constrained distributions, which has posed significant challenges to both non-asymptotic analysis and algorithmic design. We propose a unified framework, which is inspired by the classical mirror descent, to derive novel first-order sampling schemes. We prove that, for a general target distribution with strongly convex potential, our framework implies the existence of a first-order algorithm achieving $\tilde{O}(ε^{-2}d)$ convergence, suggesting that the state-of-the-art $\tilde{O}(ε^{-6}d^5)$ can be vastly improved. With the important Latent Dirichlet Allocation (LDA) application in mind, we specialize our algorithm to sample from Dirichlet posteriors, and derive the first non-asymptotic $\tilde{O}(ε^{-2}d^2)$ rate for first-order sampling. We further extend our framework to the mini-batch setting and prove convergence rates when only stochastic gradients are available. Finally, we report promising experimental results for LDA on real datasets.

cs.LG↗

arXiv2020

Robust Reinforcement Learning via Adversarial training with Langevin Dynamics

We introduce a sampling perspective to tackle the challenging task of training robust Reinforcement Learning (RL) agents. Leveraging the powerful Stochastic Gradient Langevin Dynamics, we present a novel, scalable two-player RL algorithm, which is a sampling variant of the two-player policy gradient method. Our algorithm consistently outperforms existing baselines, in terms of generalization across different training and testing conditions, on several MuJoCo environments. Our experiments also show that, even for objective functions that entirely ignore potential environmental shifts, our sampling approach remains highly robust in comparison to standard RL algorithms.

cs.LG↗

arXiv2020

Conditional gradient methods for stochastically constrained convex minimization

We propose two novel conditional gradient-based methods for solving structured stochastic convex optimization problems with a large number of linear constraints. Instances of this template naturally arise from SDP-relaxations of combinatorial problems, which involve a number of constraints that is polynomial in the problem dimension. The most important feature of our framework is that only a subset of the constraints is processed at each iteration, thus gaining a computational advantage over prior works that require full passes. Our algorithms rely on variance reduction and smoothing used in conjunction with conditional gradient steps, and are accompanied by rigorous convergence guarantees. Preliminary numerical experiments are provided for illustrating the practical performance of the methods.

cs.LG↗

arXiv2018

Finding Mixed Nash Equilibria of Generative Adversarial Networks

We reconsider the training objective of Generative Adversarial Networks (GANs) from the mixed Nash Equilibria (NE) perspective. Inspired by the classical prox methods, we develop a novel algorithmic framework for GANs via an infinite-dimensional two-player game and prove rigorous convergence rates to the mixed NE, resolving the longstanding problem that no provably convergent algorithm exists for general GANs. We then propose a principled procedure to reduce our novel prox methods to simple sampling routines, leading to practically efficient algorithms. Finally, we provide experimental evidence that our approach outperforms methods that seek pure strategy equilibria, such as SGD, Adam, and RMSProp, both in speed and quality.

cs.LG↗

arXiv2018

Let's be Honest: An Optimal No-Regret Framework for Zero-Sum Games

We revisit the problem of solving two-player zero-sum games in the decentralized setting. We propose a simple algorithmic framework that simultaneously achieves the best rates for honest regret as well as adversarial regret, and in addition resolves the open problem of removing the logarithmic terms in convergence to the value of the game. We achieve this goal in three steps. First, we provide a novel analysis of the optimistic mirror descent (OMD), showing that it can be modified to guarantee fast convergence for both honest regret and value of the game, when the players are playing collaboratively. Second, we propose a new algorithm, dubbed as robust optimistic mirror descent (ROMD), which attains optimal adversarial regret without knowing the time horizon beforehand. Finally, we propose a simple signaling scheme, which enables us to bridge OMD and ROMD to achieve the best of both worlds. Numerical examples are presented to support our theoretical claims and show that our non-adaptive ROMD algorithm can be competitive to OMD with adaptive step-size selection.

cs.GT↗

arXiv2018

Dimension-free Information Concentration via Exp-Concavity

Information concentration of probability measures have important implications in learning theory. Recently, it is discovered that the information content of a log-concave distribution concentrates around their differential entropy, albeit with an unpleasant dependence on the ambient dimension. In this work, we prove that if the potentials of the log-concave distribution are exp-concave, which is a central notion for fast rates in online and statistical learning, then the concentration of information can be further improved to depend only on the exp-concavity parameter, and hence, it can be dimension independent. Central to our proof is a novel yet simple application of the variance Brascamp-Lieb inequality. In the context of learning theory, our concentration-of-information result immediately implies high-probability results to many of the previous bounds that only hold in expectation.

cs.LG↗

arXiv2016

Nanostructure analysis of InGaN/GaN quantum wells based on semi-polar-faced GaN nanorods

We demonstrate a series of InGaN/GaN double quantum well nanostructure elements. We grow a layer of 2 μm undoped GaN template on top of a (0001)-direction sapphire substrate. A 100 nm SiO2 thin film is deposited on top as a masking pattern layer. This layer is then covered with a 300 nm aluminum layer as the anodic aluminum oxide (AAO) hole pattern layer. After oxalic acid etching, we transfer the hole pattern from the AAO layer to the SiO2 layer by reactive ion etching. Lastly, we utilize metal-organic chemical vapor deposition to grow GaN nanorods approximately 1.5 μm in size. We then grow two layers of InGaN/GaN double quantum wells on the semi-polar face of the GaN nanorod substrate under different temperatures. We then study the characteristics of the InGaN/GaN quantum wells formed on the semi-polar faces of GaN nanorods. We report the following findings from our study: first, using SiO2 with repeating hole pattern, we are able to grow high-quality GaN nanorods with diameters of approximately 80-120 nm; second, photoluminescence (PL) measurements enable us to identify Fabry-Perot effect from InGaN/GaN quantum wells on the semi-polar face. We calculate the quantum wells' cavity thickness with obtained PL measurements. Lastly, high resolution TEM images allow us to study the lattice structure characteristics of InGaN/GaN quantum wells on GaN nanorod and identify the existence of threading dislocations in the lattice structure that affects the GaN nanorod's growth mechanism.

cond-mat.mtrl-sci↗

arXiv2016

An Efficient Streaming Algorithm for the Submodular Cover Problem

We initiate the study of the classical Submodular Cover (SC) problem in the data streaming model which we refer to as the Streaming Submodular Cover (SSC). We show that any single pass streaming algorithm using sublinear memory in the size of the stream will fail to provide any non-trivial approximation guarantees for SSC. Hence, we consider a relaxed version of SSC, where we only seek to find a partial cover. We design the first Efficient bicriteria Submodular Cover Streaming (ESC-Streaming) algorithm for this problem, and provide theoretical guarantees for its performance supported by numerical evidence. Our algorithm finds solutions that are competitive with the near-optimal offline greedy algorithm despite requiring only a single pass over the data stream. In our numerical experiments, we evaluate the performance of ESC-Streaming on active set selection and large-scale graph cover problems.

cs.DS↗

arXiv2016

Frank-Wolfe Works for Non-Lipschitz Continuous Gradient Objectives: Scalable Poisson Phase Retrieval

We study a phase retrieval problem in the Poisson noise model. Motivated by the PhaseLift approach, we approximate the maximum-likelihood estimator by solving a convex program with a nuclear norm constraint. While the Frank-Wolfe algorithm, together with the Lanczos method, can efficiently deal with nuclear norm constraints, our objective function does not have a Lipschitz continuous gradient, and hence existing convergence guarantees for the Frank-Wolfe algorithm do not apply. In this paper, we show that the Frank-Wolfe algorithm works for the Poisson phase retrieval problem, and has a global convergence rate of O(1/t), where t is the iteration counter. We provide rigorous theoretical guarantee and illustrating numerical results.

math.OC↗

arXiv2015

A Geometric View on Constrained M-Estimators

We study the estimation error of constrained M-estimators, and derive explicit upper bounds on the expected estimation error determined by the Gaussian width of the constraint set. Both of the cases where the true parameter is on the boundary of the constraint set (matched constraint), and where the true parameter is strictly in the constraint set (mismatched constraint) are considered. For both cases, we derive novel universal estimation error bounds for regression in a generalized linear model with the canonical link function. Our error bound for the mismatched constraint case is minimax optimal in terms of its dependence on the sample size, for Gaussian linear regression by the Lasso.

math.ST↗

arXiv2013

Mathematical Foundations for Information Theory in Diffusion-Based Molecular Communications

Molecular communication emerges as a promising communication paradigm for nanotechnology. However, solid mathematical foundations for information-theoretic analysis of molecular communication have not yet been built. In particular, no one has ever proven that the channel coding theorem applies for molecular communication, and no relationship between information rate capacity (maximum mutual information) and code rate capacity (supremum achievable code rate) has been established. In this paper, we focus on a major subclass of molecular communication - the diffusion-based molecular communication. We provide solid mathematical foundations for information theory in diffusion-based molecular communication by creating a general diffusion-based molecular channel model in measure-theoretic form and prove its channel coding theorems. Various equivalence relationships between statistical and operational definitions of channel capacity are also established, including the most classic information rate capacity and code rate capacity. As byproducts, we have shown that the diffusion-based molecular channel is with "asymptotically decreasing input memory and anticipation" and "d-continuous". Other properties of diffusion-based molecular channel such as stationarity or ergodicity are also proven.

cs.IT↗