arXiv subjects

Nan Jiang

Publications and source records attributed to Nan Jiang.

At least 181 records · Page 10Linked to original sources

arXiv2023

Offline Learning in Markov Games with General Function Approximation

We study offline multi-agent reinforcement learning (RL) in Markov games, where the goal is to learn an approximate equilibrium -- such as Nash equilibrium and (Coarse) Correlated Equilibrium -- from an offline dataset pre-collected from the game. Existing works consider relatively restricted tabular or linear models and handle each equilibria separately. In this work, we provide the first framework for sample-efficient offline learning in Markov games under general function approximation, handling all 3 equilibria in a unified manner. By using Bellman-consistent pessimism, we obtain interval estimation for policies' returns, and use both the upper and the lower bounds to obtain a relaxation on the gap of a candidate policy, which becomes our optimization objective. Our results generalize prior works and provide several additional insights. Importantly, we require a data coverage condition that improves over the recently proposed "unilateral concentrability". Our condition allows selective coverage of deviation policies that optimally trade-off between their greediness (as approximate best responses) and coverage, and we show scenarios where this leads to significantly better guarantees. As a new connection, we also show how our algorithmic framework can subsume seemingly different solution concepts designed for the special case of two-player zero-sum games.

cs.LG↗

arXiv2023

Reinforcement Learning in Low-Rank MDPs with Density Features

MDPs with low-rank transitions -- that is, the transition matrix can be factored into the product of two matrices, left and right -- is a highly representative structure that enables tractable learning. The left matrix enables expressive function approximation for value-based learning and has been studied extensively. In this work, we instead investigate sample-efficient learning with density features, i.e., the right matrix, which induce powerful models for state-occupancy distributions. This setting not only sheds light on leveraging unsupervised learning in RL, but also enables plug-in solutions for convex RL. In the offline setting, we propose an algorithm for off-policy estimation of occupancies that can handle non-exploratory data. Using this as a subroutine, we further devise an online algorithm that constructs exploratory data distributions in a level-by-level manner. As a central technical challenge, the additive error of occupancy estimation is incompatible with the multiplicative definition of data coverage. In the absence of strong assumptions like reachability, this incompatibility easily leads to exponential error blow-up, which we overcome via novel technical tools. Our results also readily extend to the representation learning setting, when the density features are unknown and must be learned from an exponentially large candidate set.

cs.LG↗

arXiv2023

ALSO-X#: Better Convex Approximations for Distributionally Robust Chance Constrained Programs

This paper studies distributionally robust chance constrained programs (DRCCPs), where the uncertain constraints must be satisfied with at least a probability of a prespecified threshold for all probability distributions from the Wasserstein ambiguity set. As DRCCPs are often nonconvex and challenging to solve optimally, researchers have been developing various convex inner approximations. Recently, ALSO-X has been proven to outperform the conditional value-at-risk (CVaR) approximation of a regular chance constrained program when the deterministic set is convex. In this work, we relax this assumption by introducing a new ALSO-X\# method for solving DRCCPs. Namely, in the bilevel reformulations of ALSO-X and CVaR approximation, we observe that the lower-level ALSO-X is a special case of the lower-level CVaR approximation and the upper-level CVaR approximation is more restricted than the one in ALSO-X. This observation motivates us to propose the ALSO-X\#, which still resembles a bilevel formulation -- in the lower-level problem, we adopt the more general CVaR approximation, and for the upper-level one, we choose the less restricted ALSO-X. We show that ALSO-X\# can always be better than the CVaR approximation and can outperform ALSO-X under regular chance constrained programs and type $\infty-$Wasserstein ambiguity set. We also provide new sufficient conditions under which ALSO-X\# outputs an optimal solution to a DRCCP. We apply the proposed ALSO-X\# to a wireless communication problem and numerically demonstrate that the solution quality can be even better than the exact method.

math.OC↗

arXiv2022

Comparison of dynamical and kinematic reference frames via pulsar positions from timing, Gaia, and interferometric astrometry

Pulsars are special objects whose positions can be determined independently from timing, radio interferometric, and Gaia astrometry at sub-milliarcsecond (mas) precision; thus, they provide a unique way to monitor the link between dynamical and kinematic reference frames. We aimed to assess the orientation consistency between the dynamical reference frame represented by the planetary ephemeris and the kinematic reference frames constructed by Gaia and VLBI through pulsar positions. We identified 49 pulsars in Gaia Data Release 3 and 62 pulsars with very long baseline interferometry (VLBI) positions from the PSR$π$ and MSPSR$π$ projects and searched for the published timing solutions of these pulsars. We then compared pulsar positions measured by timing, VLBI, and Gaia to estimate the orientation offsets of the ephemeris frames with respect to the Gaia and VLBI reference frames by iterative fitting. We found orientation offsets of $\sim$10 mas in the DE200 frame with respect to the Gaia and VLBI frame. Our results depend strongly on the subset used in the comparison and could be biased by underestimated errors in the archival timing data, reflecting the limitation of using the literature timing solutions to determine the frame rotation.

astro-ph.IM↗

arXiv2022

Toward Multi-Service Edge-Intelligence Paradigm: Temporal-Adaptive Prediction for Time-Critical Control over Wireless

Time-critical control applications typically pose stringent connectivity requirements for communication networks. The imperfections associated with the wireless medium such as packet losses, synchronization errors, and varying delays have a detrimental effect on performance of real-time control, often with safety implications. This paper introduces multi-service edge-intelligence as a new paradigm for realizing time-critical control over wireless. It presents the concept of multi-service edge-intelligence which revolves around tight integration of wireless access, edge-computing and machine learning techniques, in order to provide stability guarantees under wireless imperfections. The paper articulates some of the key system design aspects of multi-service edge-intelligence. It also presents a temporal-adaptive prediction technique to cope with dynamically changing wireless environments. It provides performance results in a robotic teleoperation scenario. Finally, it discusses some open research and design challenges for multi-service edge-intelligence.

cs.NI↗

arXiv2022

ARMOR: A Model-based Framework for Improving Arbitrary Baseline Policies with Offline Data

We propose a new model-based offline RL framework, called Adversarial Models for Offline Reinforcement Learning (ARMOR), which can robustly learn policies to improve upon an arbitrary baseline policy regardless of data coverage. Based on the concept of relative pessimism, ARMOR is designed to optimize for the worst-case relative performance when facing uncertainty. In theory, we prove that the learned policy of ARMOR never degrades the performance of the baseline policy with any admissible hyperparameter, and can learn to compete with the best policy within data coverage when the hyperparameter is well tuned, and the baseline policy is supported by the data. Such a robust policy improvement property makes ARMOR especially suitable for building real-world learning systems, because in practice ensuring no performance degradation is imperative before considering any benefit learning can bring.

cs.LG↗

arXiv2022

Beyond the Return: Off-policy Function Estimation under User-specified Error-measuring Distributions

Off-policy evaluation often refers to two related tasks: estimating the expected return of a policy and estimating its value function (or other functions of interest, such as density ratios). While recent works on marginalized importance sampling (MIS) show that the former can enjoy provable guarantees under realizable function approximation, the latter is only known to be feasible under much stronger assumptions such as prohibitively expressive discriminators. In this work, we provide guarantees for off-policy function estimation under only realizability, by imposing proper regularization on the MIS objectives. Compared to commonly used regularization in MIS, our regularizer is much more flexible and can account for an arbitrary user-specified distribution, under which the learned function will be close to the groundtruth. We provide exact characterization of the optimal dual solution that needs to be realized by the discriminator class, which determines the data-coverage assumption in the case of value-function learning. As another surprising observation, the regularizer can be altered to relax the data-coverage requirement, and completely eliminate it in the ideal case with strong side information.

cs.LG↗

arXiv2022

On the Statistical Efficiency of Reward-Free Exploration in Non-Linear RL

We study reward-free reinforcement learning (RL) under general non-linear function approximation, and establish sample efficiency and hardness results under various standard structural assumptions. On the positive side, we propose the RFOLIVE (Reward-Free OLIVE) algorithm for sample-efficient reward-free exploration under minimal structural assumptions, which covers the previously studied settings of linear MDPs (Jin et al., 2020b), linear completeness (Zanette et al., 2020b) and low-rank MDPs with unknown representation (Modi et al., 2021). Our analyses indicate that the explorability or reachability assumptions, previously made for the latter two settings, are not necessary statistically for reward-free exploration. On the negative side, we provide a statistical hardness result for both reward-free and reward-aware exploration under linear completeness assumptions when the underlying features are unknown, showing an exponential separation between low-rank and linear completeness settings.

cs.LG↗

arXiv2022

Three linear, unconditionally stable, second order decoupling methods for the Allen--Cahn--Navier--Stokes phase field model

Hydrodynamics coupled phase field models have intricate difficulties to solve numerically as they feature high nonlinearity and great complexity in coupling. In this paper, we propose three second order, linear, unconditionally stable decoupling methods based on the Crank--Nicolson leap-frog time discretization for solving the Allen--Cahn--Navier--Stokes (ACNS) phase field model of two-phase incompressible flows. The ACNS system is decoupled via the artificial compression method and a splitting approach by introducing an exponential scalar auxiliary variable.We prove all three algorithms are unconditionally long time stable. Numerical examples are provided to verify the convergence rate, unconditional stability, and computational efficiency.

math.NA↗

arXiv2022

Interaction-Grounded Learning with Action-inclusive Feedback

Consider the problem setting of Interaction-Grounded Learning (IGL), in which a learner's goal is to optimally interact with the environment with no explicit reward to ground its policies. The agent observes a context vector, takes an action, and receives a feedback vector, using this information to effectively optimize a policy with respect to a latent reward function. Prior analyzed approaches fail when the feedback vector contains the action, which significantly limits IGL's success in many potential scenarios such as Brain-computer interface (BCI) or Human-computer interface (HCI) applications. We address this by creating an algorithm and analysis which allows IGL to work even when the feedback vector contains the action, encoded in any fashion. We provide theoretical guarantees and large-scale experiments based on supervised datasets to demonstrate the effectiveness of the new approach.

cs.LG↗

arXiv2022

The Role of Coverage in Online Reinforcement Learning

Coverage conditions -- which assert that the data logging distribution adequately covers the state space -- play a fundamental role in determining the sample complexity of offline reinforcement learning. While such conditions might seem irrelevant to online reinforcement learning at first glance, we establish a new connection by showing -- somewhat surprisingly -- that the mere existence of a data distribution with good coverage can enable sample-efficient online RL. Concretely, we show that coverability -- that is, existence of a data distribution that satisfies a ubiquitous coverage condition called concentrability -- can be viewed as a structural property of the underlying MDP, and can be exploited by standard algorithms for sample-efficient exploration, even when the agent does not know said distribution. We complement this result by proving that several weaker notions of coverage, despite being sufficient for offline RL, are insufficient for online RL. We also show that existing complexity measures for online RL, including Bellman rank and Bellman-Eluder dimension, fail to optimally capture coverability, and propose a new complexity measure, the sequential extrapolation coefficient, to provide a unification.

cs.LG↗

arXiv2022

Second order, unconditionally stable, linear ensemble algorithms for the magnetohydrodynamics equations

We propose two unconditionally stable, linear ensemble algorithms with pre-computable shared coefficient matrices across different realizations for the magnetohydrodynamics equations. The viscous terms are treated by a standard perturbative discretization. The nonlinear terms are discretized fully explicitly within the framework of the generalized positive auxiliary variable approach (GPAV). Artificial viscosity stabilization that modifies the kinetic energy is introduced to improve accuracy of the GPAV ensemble methods. Numerical results are presented to demonstrate the accuracy and robustness of the ensemble algorithms.

math.NA↗

arXiv2022

Towards Deployment-Efficient Reinforcement Learning: Lower Bound and Optimality

Deployment efficiency is an important criterion for many real-world applications of reinforcement learning (RL). Despite the community's increasing interest, there lacks a formal theoretical formulation for the problem. In this paper, we propose such a formulation for deployment-efficient RL (DE-RL) from an "optimization with constraints" perspective: we are interested in exploring an MDP and obtaining a near-optimal policy within minimal \emph{deployment complexity}, whereas in each deployment the policy can sample a large batch of data. Using finite-horizon linear MDPs as a concrete structural model, we reveal the fundamental limit in achieving deployment efficiency by establishing information-theoretic lower bounds, and provide algorithms that achieve the optimal deployment efficiency. Moreover, our formulation for DE-RL is flexible and can serve as a building block for other practically relevant settings; we give "Safe DE-RL" and "Sample-Efficient DE-RL" as two examples, which may be worth future investigation.

cs.LG↗

arXiv2022

Effect of twin boundaries on the strength of body-centered cubic tungsten nanowires

Twin boundaries (TBs) are assumed to be obstacles to dislocation motion and increase the strength of metals. Here, we report the abnormal phenomenon that TBs reduce the strength of body-centered cubic (BCC) tungsten (W). [1-11]-oriented W nanowires with (121) twin planes and free of dislocations were fabricated by chemical vapor deposition. In situ tensile tests within the transmission electron microscope were performed on single-crystal and twinned W nanowires. The fracture strength of the twinned W nanowire was 13.7 GPa, 16% lower than that of the single-crystal W nanowire (16.3 GPa). The weakening mechanism of the TB was revealed by a combination of atomic-resolution characterizations and atomistic simulations. Twinned W nanowires failed by the early nucleation of a crack at the intersection of the TB with the surface. The standard strengthening mechanism by dislocation/TB interaction was not operative in W because the high Peierls barrier and stacking fault energy in W hinder dislocation nucleation and glide. These findings provide a new insight into the influence of TBs on the mechanical properties of BCC metals.

cond-mat.mtrl-sci↗

arXiv2022

On the Value of Behavioral Representations for Dense Retrieval

We consider text retrieval within dense representational space in real-world settings such as e-commerce search where (a) document popularity and (b) diversity of queries associated with a document have a skewed distribution. Most of the contemporary dense retrieval literature presents two shortcomings in these settings. (1) They learn an almost equal number of representations per document, agnostic to the fact that a few head documents are disproportionately more critical to achieving a good retrieval performance. (ii) They learn purely semantic document representations inferred from intrinsic document characteristics which may not contain adequate information to determine the queries for which the document is relevant--especially when the document is short. We propose to overcome these limitations by augmenting semantic document representations learned by bi-encoders with behavioral document representations learned by our proposed approach MVG. To do so, MVG (1) determines how to divide the total budget for behavioral representations by drawing a connection to the Pitman-Yor process, and (2) simply clusters the queries related to a given document (based on user behavior) within the representational space learned by a base bi-encoder, and treats the cluster centers as its behavioral representations. Our central contribution is the finding such a simple intuitive light-weight approach leads to substantial gains in key first-stage retrieval metrics by incurring only a marginal memory overhead. We establish this via extensive experiments over three large public datasets comparing several single-vector and multi-vector bi-encoders, a proprietary e-commerce search dataset compared to production-quality bi-encoder, and an A/B test.

cs.IR↗

arXiv2022

Finite Sample Analysis of Minimax Offline Reinforcement Learning: Completeness, Fast Rates and First-Order Efficiency

We offer a theoretical characterization of off-policy evaluation (OPE) in reinforcement learning using function approximation for marginal importance weights and $q$-functions when these are estimated using recent minimax methods. Under various combinations of realizability and completeness assumptions, we show that the minimax approach enables us to achieve a fast rate of convergence for weights and quality functions, characterized by the critical inequality \citep{bartlett2005}. Based on this result, we analyze convergence rates for OPE. In particular, we introduce novel alternative completeness conditions under which OPE is feasible and we present the first finite-sample result with first-order efficiency in non-tabular environments, i.e., having the minimal coefficient in the leading term.

cs.LG↗

arXiv2022

A Few Expert Queries Suffices for Sample-Efficient RL with Resets and Linear Value Approximation

The current paper studies sample-efficient Reinforcement Learning (RL) in settings where only the optimal value function is assumed to be linearly-realizable. It has recently been understood that, even under this seemingly strong assumption and access to a generative model, worst-case sample complexities can be prohibitively (i.e., exponentially) large. We investigate the setting where the learner additionally has access to interactive demonstrations from an expert policy, and we present a statistically and computationally efficient algorithm (Delphi) for blending exploration with expert queries. In particular, Delphi requires $\tilde{\mathcal{O}}(d)$ expert queries and a $\texttt{poly}(d,H,|\mathcal{A}|,1/\varepsilon)$ amount of exploratory samples to provably recover an $\varepsilon$-suboptimal policy. Compared to pure RL approaches, this corresponds to an exponential improvement in sample complexity with surprisingly-little expert input. Compared to prior imitation learning (IL) approaches, our required number of expert demonstrations is independent of $H$ and logarithmic in $1/\varepsilon$, whereas all prior work required at least linear factors of both in addition to the same dependence on $d$. Towards establishing the minimal amount of expert queries needed, we show that, in the same setting, any learner whose exploration budget is polynomially-bounded (in terms of $d,H,$ and $|\mathcal{A}|$) will require at least $\tildeΩ(\sqrt{d})$ oracle calls to recover a policy competing with the expert's value function. Under the weaker assumption that the expert's policy is linear, we show that the lower bound increases to $\tildeΩ(d)$.

cs.LG↗

arXiv2022

Adversarially Trained Actor Critic for Offline Reinforcement Learning

We propose Adversarially Trained Actor Critic (ATAC), a new model-free algorithm for offline reinforcement learning (RL) under insufficient data coverage, based on the concept of relative pessimism. ATAC is designed as a two-player Stackelberg game: A policy actor competes against an adversarially trained value critic, who finds data-consistent scenarios where the actor is inferior to the data-collection behavior policy. We prove that, when the actor attains no regret in the two-player game, running ATAC produces a policy that provably 1) outperforms the behavior policy over a wide range of hyperparameters that control the degree of pessimism, and 2) competes with the best policy covered by data with appropriately chosen hyperparameters. Compared with existing works, notably our framework offers both theoretical guarantees for general function approximation and a deep RL implementation scalable to complex environments and large datasets. In the D4RL benchmark, ATAC consistently outperforms state-of-the-art offline RL algorithms on a range of continuous control tasks.

cs.LG↗