arXiv subjects

Qiang Liu

Publications and source records attributed to Qiang Liu.

At least 469 records · Page 26Linked to original sources

arXiv2018

Off-Policy Evaluation and Learning from Logged Bandit Feedback: Error Reduction via Surrogate Policy

When learning from a batch of logged bandit feedback, the discrepancy between the policy to be learned and the off-policy training data imposes statistical and computational challenges. Unlike classical supervised learning and online learning settings, in batch contextual bandit learning, one only has access to a collection of logged feedback from the actions taken by a historical policy, and expect to learn a policy that takes good actions in possibly unseen contexts. Such a batch learning setting is ubiquitous in online and interactive systems, such as ad platforms and recommendation systems. Existing approaches based on inverse propensity weights, such as Inverse Propensity Scoring (IPS) and Policy Optimizer for Exponential Models (POEM), enjoy unbiasedness but often suffer from large mean squared error. In this work, we introduce a new approach named Maximum Likelihood Inverse Propensity Scoring (MLIPS) for batch learning from logged bandit feedback. Instead of using the given historical policy as the proposal in inverse propensity weights, we estimate a maximum likelihood surrogate policy based on the logged action-context pairs, and then use this surrogate policy as the proposal. We prove that MLIPS is asymptotically unbiased, and moreover, has a smaller nonasymptotic mean squared error than IPS. Such an error reduction phenomenon is somewhat surprising as the estimated surrogate policy is less accurate than the given historical policy. Results on multi-label classification problems and a large- scale ad placement dataset demonstrate the empirical effectiveness of MLIPS. Furthermore, the proposed surrogate policy technique is complementary to existing error reduction techniques, and when combined, is able to consistently boost the performance of several widely used approaches.

cs.LG↗

arXiv2018

Stein Variational Message Passing for Continuous Graphical Models

We propose a novel distributed inference algorithm for continuous graphical models, by extending Stein variational gradient descent (SVGD) to leverage the Markov dependency structure of the distribution of interest. Our approach combines SVGD with a set of structured local kernel functions defined on the Markov blanket of each node, which alleviates the curse of high dimensionality and simultaneously yields a distributed algorithm for decentralized inference tasks. We justify our method with theoretical analysis and show that the use of local kernels can be viewed as a new type of localized approximation that matches the target distribution on the conditional distributions of each node over its Markov blanket. Our empirical results show that our method outperforms a variety of baselines including standard MCMC and particle message passing methods.

stat.ML↗

arXiv2018

Stein Variational Gradient Descent Without Gradient

Stein variational gradient decent (SVGD) has been shown to be a powerful approximate inference algorithm for complex distributions. However, the standard SVGD requires calculating the gradient of the target density and cannot be applied when the gradient is unavailable. In this work, we develop a gradient-free variant of SVGD (GF-SVGD), which replaces the true gradient with a surrogate gradient, and corrects the induced bias by re-weighting the gradients in a proper form. We show that our GF-SVGD can be viewed as the standard SVGD with a special choice of kernel, and hence directly inherits the theoretical properties of SVGD. We shed insights on the empirical choice of the surrogate gradient and propose an annealed GF-SVGD that leverages the idea of simulated annealing to improve the performance on high dimensional complex distributions. Empirical studies show that our method consistently outperforms a number of recent advanced gradient-free MCMC methods.

stat.ML↗

arXiv2018

The First Release of the AST3-1 Point Source Catalogue from Dome A, Antarctica

The three Antarctic Survey Telescopes (AST3) aim to carry out time domain imaging survey at Dome A, Antarctica. The first of the three telescopes (AST3-1) was successfully deployed on January 2012. AST3-1 is a 500\,mm aperture modified Schmidt telescope with a 680\,mm diameter primary mirror. AST3-1 is equipped with a SDSS $i$ filter and a 10k $\times$ 10k frame transfer CCD camera, reduced to 5k $\times$ 10k by electronic shuttering, resulting in a 4.3 deg$^2$ field-of-view. To verify the capability of AST3-1 for a variety of science goals, extensive commissioning was carried out between March and May 2012. The commissioning included a survey covering 2000 deg$^2$ as well as the entire Large and Small Magellanic Clouds. Frequent repeated images were made of the center of the Large Magellanic Cloud, a selected exoplanet transit field, and fields including some Wolf-Rayet stars. Here we present the data reduction and photometric measurements of the point sources observed by AST3-1. We have achieved a survey depth of 19.3\,mag in 60 s exposures with 5\,mmag precision in the light curves of bright stars. The facility achieves sub-mmag photometric precision under stable survey conditions, approaching its photon noise limit. These results demonstrate that AST3-1 at Dome A is extraordinarily competitive in time-domain astronomy, including both quick searches for faint transients and the detection of tiny transit signals.

astro-ph.IM↗

arXiv2018

Investigation of the near-threshold cluster resonance in $^{14}\rm{C}$

An experiment for $p(^{14}\rm{C}$,$^{14}\rm{C}^{*}\rightarrow^{10}\rm{Be}+α)\mathit{p}$ inelastic excitation and decay was performed in inverse kinematics at a beam energy of 25.3 MeV/u. A series of $^{14}\rm{C}$ excited states, including a new one at 18.3(1) MeV, were observed which decay to various states of the final nucleus of $^{10}\rm{Be}$. A specially designed telescope-system, installed around the zero degree, played an essential role in detecting the resonant states near the $α$-separation threshold. A state at 14.1(1) MeV is clearly identified, being consistent with the predicted band-head of the molecular rotational band characterized by the $π$-bond linear-chain-configuration. Further clarification of the properties of this exotic state is suggested by using appropriate reaction tools.

nucl-ex↗

arXiv2018

Topological Maxwell Metal Bands in a Superconducting Qutrit

We experimentally explore the topological Maxwell metal bands by mapping the momentum space of condensed-matter models to the tunable parameter space of superconducting quantum circuits. An exotic band structure that is effectively described by the spin-1 Maxwell equations is imaged. Three-fold degenerate points dubbed Maxwell points are observed in the Maxwell metal bands. Moreover, we engineer and observe the topological phase transition from the topological Maxwell metal to a trivial insulator, and report the first experiment to measure the Chern numbers that are higher than one.

quant-ph↗

arXiv2018

Stochastic Variance Reduction for Policy Gradient Estimation

Recent advances in policy gradient methods and deep learning have demonstrated their applicability for complex reinforcement learning problems. However, the variance of the performance gradient estimates obtained from the simulation is often excessive, leading to poor sample efficiency. In this paper, we apply the stochastic variance reduced gradient descent (SVRG) to model-free policy gradient to significantly improve the sample-efficiency. The SVRG estimation is incorporated into a trust-region Newton conjugate gradient framework for the policy optimization. On several Mujoco tasks, our method achieves significantly better performance compared to the state-of-the-art model-free policy gradient methods in robotic continuous control such as trust region policy optimization (TRPO)

cs.LG↗

arXiv2018

Learning to Explore with Meta-Policy Gradient

The performance of off-policy learning, including deep Q-learning and deep deterministic policy gradient (DDPG), critically depends on the choice of the exploration policy. Existing exploration methods are mostly based on adding noise to the on-going actor policy and can only explore \emph{local} regions close to what the actor policy dictates. In this work, we develop a simple meta-policy gradient algorithm that allows us to adaptively learn the exploration policy in DDPG. Our algorithm allows us to train flexible exploration behaviors that are independent of the actor policy, yielding a \emph{global exploration} that significantly speeds up the learning process. With an extensive study, we show that our method significantly improves the sample-efficiency of DDPG on a variety of reinforcement learning tasks.

cs.LG↗

arXiv2018

Emulating topological chiral magnetic effects in artificial Weyl semimetals

We realized highly tunable Weyl semimetal-bands and subsequently emulated the topological chiral magnetic effects in superconducting quantum circuits. Driving the superconducting quantum circuits with elaborately designed microwave fields, we mapped the momentum space of a lattice to the parameter space, realizing the Hamiltonian of a Weyl semimetal. By measuring the energy spectrum, we directly imaged the Weyl points of cubic lattices, whose topological winding numbers were further determined from the Berry curvature measurement. In particular, we used an additional microwave field to produce a momentum-dependent chemical potential, from which the chiral magnetic topological current was extracted in the presence of an artificial magnetic field. This pure topological current is proportional to the magnetic field, which is in contrast to the famous Ampere's law, and may have significant impacts on topological materials and quantum devices.

cond-mat.mes-hall↗

arXiv2018

BEBP: An Poisoning Method Against Machine Learning Based IDSs

In big data era, machine learning is one of fundamental techniques in intrusion detection systems (IDSs). However, practical IDSs generally update their decision module by feeding new data then retraining learning models in a periodical way. Hence, some attacks that comprise the data for training or testing classifiers significantly challenge the detecting capability of machine learning-based IDSs. Poisoning attack, which is one of the most recognized security threats towards machine learning-based IDSs, injects some adversarial samples into the training phase, inducing data drifting of training data and a significant performance decrease of target IDSs over testing data. In this paper, we adopt the Edge Pattern Detection (EPD) algorithm to design a novel poisoning method that attack against several machine learning algorithms used in IDSs. Specifically, we propose a boundary pattern detection algorithm to efficiently generate the points that are near to abnormal data but considered to be normal ones by current classifiers. Then, we introduce a Batch-EPD Boundary Pattern (BEBP) detection algorithm to overcome the limitation of the number of edge pattern points generated by EPD and to obtain more useful adversarial samples. Based on BEBP, we further present a moderate but effective poisoning method called chronic poisoning attack. Extensive experiments on synthetic and three real network data sets demonstrate the performance of the proposed poisoning method against several well-known machine learning algorithms and a practical intrusion detection method named FMIFS-LSSVM-IDS.

cs.LG↗

arXiv2018

Action-depedent Control Variates for Policy Optimization via Stein's Identity

Policy gradient methods have achieved remarkable successes in solving challenging reinforcement learning problems. However, it still often suffers from the large variance issue on policy gradient estimation, which leads to poor sample efficiency during training. In this work, we propose a control variate method to effectively reduce variance for policy gradient methods. Motivated by the Stein's identity, our method extends the previous control variate methods used in REINFORCE and advantage actor-critic by introducing more general action-dependent baseline functions. Empirical studies show that our method significantly improves the sample efficiency of the state-of-the-art policy gradient approaches.

stat.ML↗

arXiv2018

On the Discrimination-Generalization Tradeoff in GANs

Generative adversarial training can be generally understood as minimizing certain moment matching loss defined by a set of discriminator functions, typically neural networks. The discriminator set should be large enough to be able to uniquely identify the true distribution (discriminative), and also be small enough to go beyond memorizing samples (generalizable). In this paper, we show that a discriminator set is guaranteed to be discriminative whenever its linear span is dense in the set of bounded continuous functions. This is a very mild condition satisfied even by neural networks with a single neuron. Further, we develop generalization bounds between the learned distribution and true distribution under different evaluation metrics. When evaluated with neural distance, our bounds show that generalization is guaranteed as long as the discriminator set is small enough, regardless of the size of the generator or hypothesis set. When evaluated with KL divergence, our bound provides an explanation on the counter-intuitive behaviors of testing likelihood in GAN training. Our analysis sheds lights on understanding the practical performance of GANs.

cs.LG↗

arXiv2018

Adaptive Scan Gibbs Sampler for Large Scale Inference Problems

For large scale on-line inference problems the update strategy is critical for performance. We derive an adaptive scan Gibbs sampler that optimizes the update frequency by selecting an optimum mini-batch size. We demonstrate performance of our adaptive batch-size Gibbs sampler by comparing it against the collapsed Gibbs sampler for Bayesian Lasso, Dirichlet Process Mixture Models (DPMM) and Latent Dirichlet Allocation (LDA) graphical models.

stat.ML↗

arXiv2017

Stein Variational Gradient Descent as Gradient Flow

Stein variational gradient descent (SVGD) is a deterministic sampling algorithm that iteratively transports a set of particles to approximate given distributions, based on an efficient gradient-based update that guarantees to optimally decrease the KL divergence within a function space. This paper develops the first theoretical analysis on SVGD, discussing its weak convergence properties and showing that its asymptotic behavior is captured by a gradient flow of the KL divergence functional under a new metric structure induced by Stein operator. We also provide a number of results on Stein operator and Stein's identity using the notion of weak derivative, including a new proof of the distinguishability of Stein discrepancy under weak conditions.

stat.ML↗

arXiv2017

Two Methods For Wild Variational Inference

Variational inference provides a powerful tool for approximate probabilistic in- ference on complex, structured models. Typical variational inference methods, however, require to use inference networks with computationally tractable proba- bility density functions. This largely limits the design and implementation of vari- ational inference methods. We consider wild variational inference methods that do not require tractable density functions on the inference networks, and hence can be applied in more challenging cases. As an example of application, we treat stochastic gradient Langevin dynamics (SGLD) as an inference network, and use our methods to automatically adjust the step sizes of SGLD, yielding significant improvement over the hand-designed step size schemes

stat.ML↗

arXiv2017

Learning to Draw Samples with Amortized Stein Variational Gradient Descent

We propose a simple algorithm to train stochastic neural networks to draw samples from given target distributions for probabilistic inference. Our method is based on iteratively adjusting the neural network parameters so that the output changes along a Stein variational gradient direction (Liu & Wang, 2016) that maximally decreases the KL divergence with the target distribution. Our method works for any target distribution specified by their unnormalized density function, and can train any black-box architectures that are differentiable in terms of the parameters we want to adapt. We demonstrate our method with a number of applications, including variational autoencoder (VAE) with expressive encoders to model complex latent space structures, and hyper-parameter learning of MCMC samplers that allows Bayesian inference to adaptively improve itself when seeing more data.

stat.ML↗

arXiv2017

Efficient Localized Inference for Large Graphical Models

We propose a new localized inference algorithm for answering marginalization queries in large graphical models with the correlation decay property. Given a query variable and a large graphical model, we define a much smaller model in a local region around the query variable in the target model so that the marginal distribution of the query variable can be accurately approximated. We introduce two approximation error bounds based on the Dobrushin's comparison theorem and apply our bounds to derive a greedy expansion algorithm that efficiently guides the selection of neighbor nodes for localized inference. We verify our theoretical bounds on various datasets and demonstrate that our localized inference algorithm can provide fast and accurate approximation for large graphical models.

stat.ML↗

arXiv2017

Learning Infinite RBMs with Frank-Wolfe

In this work, we propose an infinite restricted Boltzmann machine~(RBM), whose maximum likelihood estimation~(MLE) corresponds to a constrained convex optimization. We consider the Frank-Wolfe algorithm to solve the program, which provides a sparse solution that can be interpreted as inserting a hidden unit at each iteration, so that the optimization process takes the form of a sequence of finite models of increasing complexity. As a side benefit, this can be used to easily and efficiently identify an appropriate number of hidden units during the optimization. The resulting model can also be used as an initialization for typical state-of-the-art RBM training algorithms such as contrastive divergence, leading to models with consistently higher test likelihood than random initialization.

cs.LG↗