BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Nan Jiang

Nan Jiang contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
Machine LearningArtificial IntelligenceComputer Visioneess.SPquant-phComputation and LanguageNetworking and Internet Architecturephysics.atom-phphysics.opticsRoboticsComputational ComplexityComputer Science and Game Theorycond-mat.mes-halleess.ASeess.SYGraphicsHuman-Computer InteractionInformation RetrievalInformation TheoryLogic in Computer Sciencemath.NANumerical AnalysisSystems and Control

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
40works
0followers
23topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

40 published item(s)

preprint2026arXiv

Awaking Spatial Intelligence in Unified Multimodal Understanding and Generation

We present JoyAI-Image, a unified multimodal foundation model for visual understanding, text-to-image generation, and instruction-guided image editing. JoyAI-Image couples a spatially enhanced Multimodal Large Language Model (MLLM) with a Multimodal Diffusion Transformer (MMDiT), allowing perception and generation to interact through a shared multimodal interface. Around this architecture, we build a scalable training recipe that combines unified instruction tuning, long-text rendering supervision, spatially grounded data, and both general and spatial editing signals. This design gives the model broad multimodal capability while strengthening geometry-aware reasoning and controllable visual synthesis. Experiments across understanding, generation, long-text rendering, and editing benchmarks show that JoyAI-Image achieves state-of-the-art or highly competitive performance. More importantly, the bidirectional loop between enhanced understanding, controllable spatial editing, and novel-view-assisted reasoning enables the model to move beyond general visual competence toward stronger spatial intelligence. These results suggest a promising path for unified visual models in downstream applications such as vision-language-action systems and world models.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

No Action Without a NOD: A Heterogeneous Multi-Agent Architecture for Reliable Service Agents

Large language model (LLM) agents have increasingly advanced service applications, such as booking flight tickets. However, these service agents suffer from unreliability in long-horizon tasks, as they often produce policy violations, tool hallucinations, and misaligned actions, which greatly impedes their real-world deployment. To address these challenges, we propose NOD (Navigator-Operator-Director), a heterogeneous multi-agent architecture for service agents. Instead of maintaining task state implicitly in dialogue context as in prior work, we externalize a structured Global State to enable explicit task state tracking and consistent decision-making by the Navigator. Besides, we introduce selective external oversight before critical actions, allowing an independent Director agent to verify execution and intervene when necessary. As such, NOD effectively mitigates error propagation and unsafe behavior in long-horizon tasks. Experiments on $τ^2$-Bench demonstrate that NOD achieves higher task success rates and critical action precision over baselines. More importantly, NOD improves the reliability of service agents by reducing policy violations, tool hallucinations, and user-intent misalignment.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Offline Two-Player Zero-Sum Markov Games with KL Regularization

We study the problem of learning Nash equilibria in offline two-player zero-sum Markov games. While existing approaches often rely on explicit pessimism to address distribution shift, we show that KL regularization alone suffices to stabilize learning and guarantee convergence. We first introduce Regularized Offline Sequential Equilibrium (ROSE), a theoretical framework that achieves a fast $\widetilde{\mathcal{O}}(1/n)$ convergence rate under \textit{unilateral concentrability}, improving over the standard $\widetilde{\mathcal{O}}(1/\sqrt{n})$ rates in unregularized settings. We then propose Sequential Offline Self-play Mirror Descent (SOS-MD), a practical model-free algorithm based on least-squares value estimation and iterative self-play updates. We prove that the last iterate of SOS-MD attains the same $\widetilde{\mathcal{O}}(1/n)$ statistical rate up to a vanishing optimization error of order $\widetilde{\mathcal{O}}(1/\sqrt{T})$ in the number of self-play iterations $T$.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Q-MMR: Off-Policy Evaluation via Recursive Reweighting and Moment Matching

We present a novel theoretical framework, Q-MMR, for off-policy evaluation in finite-horizon MDPs. Q-MMR learns a set of scalar weights, one for each data point, such that the reweighted rewards approximate the expected return under the target policy. The weights are learned inductively in a top-down manner via a moment matching objective against a value-function discriminator class. Notably, and perhaps surprisingly, a data-dependent finite-sample guarantee for general function approximation can be established under only the realizability of $Q^π$, with a dimension-free bound -- that is, the error does not depend on the statistical complexity of the function class. We also establish connections to several existing methods, such as importance sampling and linear FQE. Further theoretical analyses shed new light on the nature of coverage, a concept of fundamental importance to offline RL.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Rethinking Importance Sampling in LLM Policy Optimization: A Cumulative Token Perspective

Reinforcement learning, including reinforcement learning with verifiable rewards (RLVR), has emerged as a powerful approach for LLM post-training. Central to these approaches is the design of the importance sampling (IS) ratio used in off-policy policy-gradient estimation. Existing methods face a fundamental bias-variance dilemma: token-level IS ratios, as adopted by PPO (Schulman et al., 2017) and GRPO (Shao et al., 2024), introduce bias by ignoring prefix state distribution mismatch; full sequence ratios provide exact trajectory-level correction but suffer from high variance due to the multiplicative accumulation of per-token ratios, while GSPO (Zheng et al., 2025) improves numerical stability via length normalization at the cost of deviating from the exact full-sequence IS correction. In this work, we identify the cumulative token IS ratio, the product of per-token ratios up to position $t$, as a theoretically principled solution to this dilemma. We prove that, under the token-level policy-gradient formulation, this ratio provides an unbiased prefix correction for each token-level gradient term and has strictly lower variance than the full sequence ratio. Building on this insight, we propose CTPO (Cumulative Token Policy Optimization), which combines the cumulative token IS ratio with position-adaptive clipping that scales log-space clip bounds according to the natural $\sqrt{t}$ growth of the cumulative log-ratio. This yields more consistent regularization across token positions. We implement and evaluate CTPO in the tool-integrated reasoning setting on several challenging mathematical reasoning benchmarks, achieving the best average performance across both model scales compared with strong GRPO and GSPO baselines. Code will be available at https://github.com/horizon-llm/CTPO.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Soft Graph Diffusion Transformer for MIMO Detection

Learning-based MIMO detection has shown strong empirical performance, yet existing methods typically rely on fixed-depth architectures without explicitly modeling the progressive refinement of symbol estimates. In this paper, we revisit MIMO detection from a flow matching perspective and propose the Soft Graph Diffusion Transformer (SGDiT), which reformulates detection as a noise-level-conditioned denoising process that progressively transforms a Gaussian initialization toward the posterior conditioned on channel observations. An adaptive layer normalization (AdaLN)-conditioned soft graph transformer is employed to parameterize the denoising dynamics, enabling stage-aware information integration between observation and symbol domains. To better align with the discrete nature of symbol detection, we further adopt a cross-entropy-based training objective that directly models bit-wise posterior probabilities, providing a more suitable inductive bias than conventional regression-based formulations. Experimental results across various MIMO system configurations demonstrate that SGDiT achieves competitive bit error rate (BER) performance compared with representative baselines. Furthermore, the proposed model exhibits good generalization capability across different channel conditions. Overall, the SGDiT framework provides an effective and practical approach for neural MIMO detection.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Spin Hall effect in van der Waals ferromagnet Fe$_{5}$GeTe$_{2}$

We investigate the spin Hall effect (SHE) in a van der Waals (vdW) ferromagnet Fe$_{5}$GeTe$_{2}$ (FGT) with a Curie temperature $T_{\rm C}$ of 310 K utilizing the spin-torque ferromagnetic resonance method. In synchronization with the emergence of the ferromagnetic phase resulting in the anomalous Hall effect (AHE), a noticeable enhancement in the SHE was observed below $T_{\rm C}$. On the other hand, the SHE shows a different temperature dependence from the AHE: the effective spin Hall conductivity is clearly enhanced with decreasing temperature unlike the anomalous Hall conductivity, reflecting the variation of band-structure accompanied by the complicated magnetic ordering of the FGT. The results provide a deep understanding of the SHE in magnetic materials to open a new route for novel functionalities in vdW materials-based spintronic devices.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Streamlining evidence based clinical recommendations with large language models

Clinical evidence underpins informed healthcare decisions, yet integrating it into real-time practice remains challenging due to intensive workloads, complex procedures, and time constraints. This study presents Quicker, an LLM-powered system that automates evidence synthesis and generates clinical recommendations following standard guideline development workflows. Quicker delivers an end-to-end pipeline from clinical questions to recommendations and supports customized decision-making through integrated tools and interactive interfaces. To evaluate how closely Quicker can reproduce guideline development processes, we constructed Q2CRBench-3, a benchmark derived from guideline development records for three diseases. Experiments show that Quicker produces precise question decomposition, expert-aligned retrieval, and near-comprehensive screening. Quicker assistance improved the accuracy of extracted study data, and its recommendations were more comprehensive and coherent than clinician-written ones. In system-level testing, Quicker working with one participant reduced recommendation development to 20-40 min. Overall, the findings demonstrate Quicker's potential to enhance the speed and reliability of evidence-based clinical decision-making.

0 citations0 reviewsNo code linkedOpen access
preprint2026arXiv

Thinking with Novel Views: A Systematic Analysis of Generative-Augmented Spatial Intelligence

Current Large Multimodal Models (LMMs) struggle with spatial reasoning tasks requiring viewpoint-dependent understanding, largely because they are confined to a single, static observation. We propose Thinking with Novel Views (TwNV), a paradigm that integrates generative novel-view synthesis into the reasoning loop: a Reasoner LMM identifies spatial ambiguity, instructs a Painter to synthesize an alternative viewpoint, and re-examines the scene with the additional evidence. Through systematic experiments we address three research questions. (1) Instruction format: numerical camera-pose specifications yield more reliable view control than free-form language. (2) Generation fidelity: synthesized view quality is tightly coupled with downstream spatial accuracy. (3) Inference-time visual scaling: iterative multi-turn view refinement further improves performance, echoing recent scaling trends in language reasoning. Across four spatial subtask categories and four LMM architectures (both closed- and open-source), TwNV consistently improves accuracy by +1.3 to +3.9 pp, with the largest gains on viewpoint-sensitive subtasks. These results establish novel-view generation as a practical lever for advancing spatial intelligence of LMMs.

0 citations0 reviewsNo code linkedOpen access
preprint2025arXiv

UniAct: Unified Motion Generation and Action Streaming for Humanoid Robots

A long-standing objective in humanoid robotics is the realization of versatile agents capable of following diverse multimodal instructions with human-level flexibility. Despite advances in humanoid control, bridging high-level multimodal perception with whole-body execution remains a significant bottleneck. Existing methods often struggle to translate heterogeneous instructions -- such as language, music, and trajectories -- into stable, real-time actions. Here we show that UniAct, a two-stage framework integrating a fine-tuned MLLM with a causal streaming pipeline, enables humanoid robots to execute multimodal instructions with sub-500 ms latency. By unifying inputs through a shared discrete codebook via FSQ, UniAct ensures cross-modal alignment while constraining motions to a physically grounded manifold. This approach yields a 19% improvement in the success rate of zero-shot tracking of imperfect reference motions. We validate UniAct on UniMoCap, our 20-hour humanoid motion benchmark, demonstrating robust generalization across diverse real-world scenarios. Our results mark a critical step toward responsive, general-purpose humanoid assistants capable of seamless interaction through unified perception and control.

0 citations0 reviewsNo code linkedOpen access
preprint2023arXiv

Solving Satisfiability Modulo Counting for Symbolic and Statistical AI Integration With Provable Guarantees

Satisfiability Modulo Counting (SMC) encompasses problems that require both symbolic decision-making and statistical reasoning. Its general formulation captures many real-world problems at the intersection of symbolic and statistical Artificial Intelligence. SMC searches for policy interventions to control probabilistic outcomes. Solving SMC is challenging because of its highly intractable nature($\text{NP}^{\text{PP}}$-complete), incorporating statistical inference and symbolic reasoning. Previous research on SMC solving lacks provable guarantees and/or suffers from sub-optimal empirical performance, especially when combinatorial constraints are present. We propose XOR-SMC, a polynomial algorithm with access to NP-oracles, to solve highly intractable SMC problems with constant approximation guarantees. XOR-SMC transforms the highly intractable SMC into satisfiability problems, by replacing the model counting in SMC with SAT formulae subject to randomized XOR constraints. Experiments on solving important SMC problems in AI for social good demonstrate that XOR-SMC finds solutions close to the true optimum, outperforming several baselines which struggle to find good approximations for the intractable model counting in SMC.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

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 $\tildeΩ(d)$.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

A Minimax Learning Approach to Off-Policy Evaluation in Confounded Partially Observable Markov Decision Processes

We consider off-policy evaluation (OPE) in Partially Observable Markov Decision Processes (POMDPs), where the evaluation policy depends only on observable variables and the behavior policy depends on unobservable latent variables. Existing works either assume no unmeasured confounders, or focus on settings where both the observation and the state spaces are tabular. In this work, we first propose novel identification methods for OPE in POMDPs with latent confounders, by introducing bridge functions that link the target policy's value and the observed data distribution. We next propose minimax estimation methods for learning these bridge functions, and construct three estimators based on these estimated bridge functions, corresponding to a value function-based estimator, a marginalized importance sampling estimator, and a doubly-robust estimator. Our proposal permits general function approximation and is thus applicable to settings with continuous or large observation/state spaces. The nonasymptotic and asymptotic properties of the proposed estimators are investigated in detail.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Adaptable Semantic Compression and Resource Allocation for Task-Oriented Communications

Task-oriented communication is a new paradigm that aims at providing efficient connectivity for accomplishing intelligent tasks rather than the reception of every transmitted bit. In this paper, a deep learning-based task-oriented communication architecture is proposed where the user extracts, compresses and transmits semantics in an end-to-end (E2E) manner. Furthermore, an approach is proposed to compress the semantics according to their importance relevant to the task, namely, adaptable semantic compression (ASC). Assuming a delay-intolerant system, supporting multiple users indicates a problem that executing with the higher compression ratio requires fewer channel resources but leads to the distortion of semantics, while executing with the lower compression ratio requires more channel resources and thus may lead to a transmission failure due to delay constraint. To solve the problem, both compression ratio and resource allocation are optimized for the task-oriented communication system to maximize the success probability of tasks. Specifically, due to the nonconvexity of the problem, we propose a compression ratio and resource allocation (CRRA) algorithm by separating the problem into two subproblems and solving iteratively to obtain the convergent solution. Furthermore, considering the scenarios where users have various service levels, a compression ratio, resource allocation, and user selection (CRRAUS) algorithm is proposed to deal with the problem. In CRRAUS, users are adaptively selected to complete the corresponding intelligent tasks based on branch and bound method at the expense of higher algorithm complexity compared with CRRA. Simulation results show that the proposed CRRA and CRRAUS algorithms can obtain at least 15% and 10% success gains over baseline algorithms, respectively.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

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.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Experimental demonstration of memory-enhanced scaling for entanglement connection of quantum repeater segments

The quantum repeater protocol is a promising approach to implement long-distance quantum communication and large-scale quantum networks. A key idea of the quantum repeater protocol is to use long-lived quantum memories to achieve efficient entanglement connection between different repeater segments with a polynomial scaling. Here we report an experiment which realizes efficient connection of two quantum repeater segments via on-demand entanglement swapping by the use of two atomic quantum memories with storage time of tens of milliseconds. With the memory enhancement, scaling-changing acceleration is demonstrated in the rate for a successful entanglement connection. The experimental realization of entanglement connection of two quantum repeater segments with an efficient memory-enhanced scaling demonstrates a key advantage of the quantum repeater protocol, which makes a cornerstone towards future large-scale quantum networks.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Model-free Representation Learning and Exploration in Low-rank MDPs

The low rank MDP has emerged as an important model for studying representation learning and exploration in reinforcement learning. With a known representation, several model-free exploration strategies exist. In contrast, all algorithms for the unknown representation setting are model-based, thereby requiring the ability to model the full dynamics. In this work, we present the first model-free representation learning algorithms for low rank MDPs. The key algorithmic contribution is a new minimax representation learning objective, for which we provide variants with differing tradeoffs in their statistical and computational properties. We interleave this representation learning step with an exploration strategy to cover the state space in a reward-free manner. The resulting algorithms are provably sample efficient and can accommodate general function approximation to scale to complex environments.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Offline Reinforcement Learning Under Value and Density-Ratio Realizability: The Power of Gaps

We consider a challenging theoretical problem in offline reinforcement learning (RL): obtaining sample-efficiency guarantees with a dataset lacking sufficient coverage, under only realizability-type assumptions for the function approximators. While the existing theory has addressed learning under realizability and under non-exploratory data separately, no work has been able to address both simultaneously (except for a concurrent work which we compare in detail). Under an additional gap assumption, we provide guarantees to a simple pessimistic algorithm based on a version space formed by marginalized importance sampling (MIS), and the guarantee only requires the data to cover the optimal policy and the function classes to realize the optimal value and density-ratio functions. While similar gap assumptions have been used in other areas of RL theory, our work is the first to identify the utility and the novel mechanism of gap assumptions in offline RL with weak function approximation.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Offline Reinforcement Learning with Realizability and Single-policy Concentrability

Sample-efficiency guarantees for offline reinforcement learning (RL) often rely on strong assumptions on both the function classes (e.g., Bellman-completeness) and the data coverage (e.g., all-policy concentrability). Despite the recent efforts on relaxing these assumptions, existing works are only able to relax one of the two factors, leaving the strong assumption on the other factor intact. As an important open problem, can we achieve sample-efficient offline RL with weak assumptions on both factors? In this paper we answer the question in the positive. We analyze a simple algorithm based on the primal-dual formulation of MDPs, where the dual variables (discounted occupancy) are modeled using a density-ratio function against offline data. With proper regularization, we show that the algorithm enjoys polynomial sample complexity, under only realizability and single-policy concentrability. We also provide alternative analyses based on different assumptions to shed light on the nature of primal-dual algorithms for offline RL.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

On the Convergence Rate of Off-Policy Policy Optimization Methods with Density-Ratio Correction

In this paper, we study the convergence properties of off-policy policy improvement algorithms with state-action density ratio correction under function approximation setting, where the objective function is formulated as a max-max-min optimization problem. We characterize the bias of the learning objective and present two strategies with finite-time convergence guarantees. In our first strategy, we present algorithm P-SREDA with convergence rate $O(ε^{-3})$, whose dependency on $ε$ is optimal. In our second strategy, we propose a new off-policy actor-critic style algorithm named O-SPIM. We prove that O-SPIM converges to a stationary point with total complexity $O(ε^{-4})$, which matches the convergence rate of some recent actor-critic algorithms in the on-policy setting.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

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.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Policy Finetuning: Bridging Sample-Efficient Offline and Online Reinforcement Learning

Recent theoretical work studies sample-efficient reinforcement learning (RL) extensively in two settings: learning interactively in the environment (online RL), or learning from an offline dataset (offline RL). However, existing algorithms and theories for learning near-optimal policies in these two settings are rather different and disconnected. Towards bridging this gap, this paper initiates the theoretical study of policy finetuning, that is, online RL where the learner has additional access to a "reference policy" $μ$ close to the optimal policy $π_\star$ in a certain sense. We consider the policy finetuning problem in episodic Markov Decision Processes (MDPs) with $S$ states, $A$ actions, and horizon length $H$. We first design a sharp offline reduction algorithm -- which simply executes $μ$ and runs offline policy optimization on the collected dataset -- that finds an $\varepsilon$ near-optimal policy within $\widetilde{O}(H^3SC^\star/\varepsilon^2)$ episodes, where $C^\star$ is the single-policy concentrability coefficient between $μ$ and $π_\star$. This offline result is the first that matches the sample complexity lower bound in this setting, and resolves a recent open question in offline RL. We then establish an $Ω(H^3S\min\{C^\star, A\}/\varepsilon^2)$ sample complexity lower bound for any policy finetuning algorithm, including those that can adaptively explore the environment. This implies that -- perhaps surprisingly -- the optimal policy finetuning algorithm is either offline reduction or a purely online RL algorithm that does not use $μ$. Finally, we design a new hybrid offline/online algorithm for policy finetuning that achieves better sample complexity than both vanilla offline reduction and purely online RL algorithms, in a relaxed setting where $μ$ only satisfies concentrability partially up to a certain time step.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Quantum Support Vector Machine without Iteration

Quantum algorithms can enhance machine learning in different aspects. In 2014, Rebentrost $et~al.$ constructed a least squares quantum support vector machine (LS-QSVM), in which the Swap Test plays a crucial role in realizing the classification. However, as the output states of a previous test cannot be reused for a new test in the Swap Test, the quantum algorithm LS-QSVM has to be repeated in preparing qubits, manipulating operations, and carrying out the measurement. This paper proposes a QSVM based on the generalized quantum amplitude estimation (AE-QSVM) which gets rid of the constraint of repetitive processes and saves the quantum resources. At first, AE-QSVM is trained by using the quantum singular value decomposition. Then, a query sample is classified by using the generalized quantum amplitude estimation in which high accuracy can be achieved by adding auxiliary qubits instead of repeating the algorithm. The complexity of AE-QSVM is reduced to $O(κ^{3}\varepsilon^{-3}(log(mn)+1))$ with an accuracy $\varepsilon$, where $m$ is the number of training vectors, $n$ is the dimension of the feature space, and $κ$ is the condition number. Experiments demonstrate that AE-QSVM is advantageous in terms of training matrix, the number of iterations, space complexity, and time complexity.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Quantum-Memory-Enhanced Preparation of Nonlocal Graph States

Graph states are an important class of multipartite entangled states. Previous experimental generation of graph states and in particular the Greenberger-Horne-Zeilinger (GHZ) states in linear optics quantum information schemes is subjected to an exponential decay in efficiency versus the system size, which limits its large-scale applications in quantum networks. Here we demonstrate an efficient scheme to prepare graph states with only a polynomial overhead using long-lived atomic quantum memories. We generate atom-photon entangled states in two atomic ensembles asynchronously, retrieve the stored atomic excitations only when both sides succeed, and further project them into a four-photon GHZ state. We measure the fidelity of this GHZ state and further demonstrate its applications in the violation of Bell-type inequalities and in quantum cryptography. Our work demonstrates the prospect of efficient generation of multipartite entangled states in large-scale distributed systems with applications in quantum information processing and metrology.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

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.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

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.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Anti-UAV: A Large Multi-Modal Benchmark for UAV Tracking

Unmanned Aerial Vehicle (UAV) offers lots of applications in both commerce and recreation. With this, monitoring the operation status of UAVs is crucially important. In this work, we consider the task of tracking UAVs, providing rich information such as location and trajectory. To facilitate research on this topic, we propose a dataset, Anti-UAV, with more than 300 video pairs containing over 580k manually annotated bounding boxes. The releasing of such a large-scale dataset could be a useful initial step in research of tracking UAVs. Furthermore, the advancement of addressing research challenges in Anti-UAV can help the design of anti-UAV systems, leading to better surveillance of UAVs. Besides, a novel approach named dual-flow semantic consistency (DFSC) is proposed for UAV tracking. Modulated by the semantic flow across video sequences, the tracker learns more robust class-level semantic information and obtains more discriminative instance-level features. Experimental results demonstrate that Anti-UAV is very challenging, and the proposed method can effectively improve the tracker's performance. The Anti-UAV benchmark and the code of the proposed approach will be publicly available at https://github.com/ucas-vg/Anti-UAV.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Minimax Model Learning

We present a novel off-policy loss function for learning a transition model in model-based reinforcement learning. Notably, our loss is derived from the off-policy policy evaluation objective with an emphasis on correcting distribution shift. Compared to previous model-based techniques, our approach allows for greater robustness under model misspecification or distribution shift induced by learning/evaluating policies that are distinct from the data-generating policy. We provide a theoretical analysis and show empirical improvements over existing model-based off-policy evaluation methods. We provide further analysis showing our loss can be used for off-policy optimization (OPO) and demonstrate its integration with more recent improvements in OPO.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

SM+: Refined Scale Match for Tiny Person Detection

Detecting tiny objects ( e.g., less than 20 x 20 pixels) in large-scale images is an important yet open problem. Modern CNN-based detectors are challenged by the scale mismatch between the dataset for network pre-training and the target dataset for detector training. In this paper, we investigate the scale alignment between pre-training and target datasets, and propose a new refined Scale Match method (termed SM+) for tiny person detection. SM+ improves the scale match from image level to instance level, and effectively promotes the similarity between pre-training and target dataset. Moreover, considering SM+ possibly destroys the image structure, a new probabilistic structure inpainting (PSI) method is proposed for the background processing. Experiments conducted across various detectors show that SM+ noticeably improves the performance on TinyPerson, and outperforms the state-of-the-art detectors with a significant margin.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

A Decoupled Learning Strategy for Massive Access Optimization in Cellular IoT Networks

Cellular-based networks are expected to offer connectivity for massive Internet of Things (mIoT) systems. However, their Random Access CHannel (RACH) procedure suffers from unreliability, due to the collision from the simultaneous massive access. Despite that this collision problem has been treated in existing RACH schemes, these schemes usually organize IoT devices' transmission and re-transmission along with fixed parameters, thus can hardly adapt to time-varying traffic patterns. Without adaptation, the RACH procedure easily suffers from high access delay, high energy consumption, or even access unavailability. With the goal of improving the RACH procedure, this paper targets to optimize the RACH procedure in real-time by maximizing a long-term hybrid multi-objective function, which consists of the number of access success devices, the average energy consumption, and the average access delay. To do so, we first optimize the long-term objective in the number of access success devices by using Deep Reinforcement Learning (DRL) algorithms for different RACH schemes, including Access Class Barring (ACB), Back-Off (BO), and Distributed Queuing (DQ). The converging capability and efficiency of different DRL algorithms including Policy Gradient (PG), Actor-Critic (AC), Deep Q-Network (DQN), and Deep Deterministic Policy Gradients (DDPG) are compared. Inspired by the results from this comparison, a decoupled learning strategy is developed to jointly and dynamically adapt the access control factors of those three access schemes. This decoupled strategy first leverage a Recurrent Neural Network (RNN) model to predict the real-time traffic values of the network environment, and then uses multiple DRL agents to cooperatively configure parameters of each RACH scheme.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

A Question Type Driven and Copy Loss Enhanced Frameworkfor Answer-Agnostic Neural Question Generation

The answer-agnostic question generation is a significant and challenging task, which aims to automatically generate questions for a given sentence but without an answer. In this paper, we propose two new strategies to deal with this task: question type prediction and copy loss mechanism. The question type module is to predict the types of questions that should be asked, which allows our model to generate multiple types of questions for the same source sentence. The new copy loss enhances the original copy mechanism to make sure that every important word in the source sentence has been copied when generating questions. Our integrated model outperforms the state-of-the-art approach in answer-agnostic question generation, achieving a BLEU-4 score of 13.9 on SQuAD. Human evaluation further validates the high quality of our generated questions. We will make our code public available for further research.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Analysis of Random Access in NB-IoT Networks with Three Coverage Enhancement Groups: A Stochastic Geometry Approach

NarrowBand-Internet of Things (NB-IoT) is a new 3GPP radio access technology designed to provide better coverage for Low Power Wide Area (LPWA) networks. To provide reliable connections with extended coverage, a repetition transmission scheme and up to three Coverage Enhancement (CE) groups are introduced into NB-IoT during both Random Access CHannel (RACH) procedure and data transmission procedure, where each CE group is configured with different repetition values and transmission resources. To characterize the RACH performance of the NB-IoT network with three CE groups, this paper develops a novel traffic-aware spatio-temporal model to analyze the RACH success probability, where both the preamble transmission outage and the collision events of each CE group jointly determine the traffic evolution and the RACH success probability. Based on this analytical model, we derive the analytical expression for the RACH success probability of a randomly chosen IoT device in each CE group over multiple time slots with different RACH schemes, including baseline, back-off (BO), access class barring (ACB), and hybrid ACB and BO schemes (ACB&BO). Our results have shown that the RACH success probabilities of the devices in three CE groups outperform that of a single CE group network but not for all the groups, which is affected by the choice of the categorizing parameters.This mathematical model and analytical framework can be applied to evaluate the performance of multiple group users of other networks with spatial separations.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

From Importance Sampling to Doubly Robust Policy Gradient

We show that on-policy policy gradient (PG) and its variance reduction variants can be derived by taking finite difference of function evaluations supplied by estimators from the importance sampling (IS) family for off-policy evaluation (OPE). Starting from the doubly robust (DR) estimator (Jiang & Li, 2016), we provide a simple derivation of a very general and flexible form of PG, which subsumes the state-of-the-art variance reduction technique (Cheng et al., 2019) as its special case and immediately hints at further variance reduction opportunities overlooked by existing literature. We analyze the variance of the new DR-PG estimator, compare it to existing methods as well as the Cramer-Rao lower bound of policy gradient, and empirically show its effectiveness.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

On Value Functions and the Agent-Environment Boundary

When function approximation is deployed in reinforcement learning (RL), the same problem may be formulated in different ways, often by treating a pre-processing step as a part of the environment or as part of the agent. As a consequence, fundamental concepts in RL, such as (optimal) value functions, are not uniquely defined as they depend on where we draw this agent-environment boundary, causing problems in theoretical analyses that provide optimality guarantees. We address this issue via a simple and novel boundary-invariant analysis of Fitted Q-Iteration, a representative RL algorithm, where the assumptions and the guarantees are invariant to the choice of boundary. We also discuss closely related issues on state resetting and Monte-Carlo Tree Search, deterministic vs stochastic systems, imitation learning, and the verifiability of theoretical assumptions from data.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Provably Efficient Q-Learning with Low Switching Cost

We take initial steps in studying PAC-MDP algorithms with limited adaptivity, that is, algorithms that change its exploration policy as infrequently as possible during regret minimization. This is motivated by the difficulty of running fully adaptive algorithms in real-world applications (such as medical domains), and we propose to quantify adaptivity using the notion of local switching cost. Our main contribution, Q-Learning with UCB2 exploration, is a model-free algorithm for H-step episodic MDP that achieves sublinear regret whose local switching cost in K episodes is $O(H^3SA\log K)$, and we provide a lower bound of $Ω(HSA)$ on the local switching cost for any no-regret algorithm. Our algorithm can be naturally adapted to the concurrent setting, which yields nontrivial results that improve upon prior work in certain aspects.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Q* Approximation Schemes for Batch Reinforcement Learning: A Theoretical Comparison

We prove performance guarantees of two algorithms for approximating $Q^\star$ in batch reinforcement learning. Compared to classical iterative methods such as Fitted Q-Iteration---whose performance loss incurs quadratic dependence on horizon---these methods estimate (some forms of) the Bellman error and enjoy linear-in-horizon error propagation, a property established for the first time for algorithms that rely solely on batch data and output stationary policies. One of the algorithms uses a novel and explicit importance-weighting correction to overcome the infamous "double sampling" difficulty in Bellman error estimation, and does not use any squared losses. Our analyses reveal its distinct characteristics and potential advantages compared to classical algorithms.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

RL-Duet: Online Music Accompaniment Generation Using Deep Reinforcement Learning

This paper presents a deep reinforcement learning algorithm for online accompaniment generation, with potential for real-time interactive human-machine duet improvisation. Different from offline music generation and harmonization, online music accompaniment requires the algorithm to respond to human input and generate the machine counterpart in a sequential order. We cast this as a reinforcement learning problem, where the generation agent learns a policy to generate a musical note (action) based on previously generated context (state). The key of this algorithm is the well-functioning reward model. Instead of defining it using music composition rules, we learn this model from monophonic and polyphonic training data. This model considers the compatibility of the machine-generated note with both the machine-generated context and the human-generated context. Experiments show that this algorithm is able to respond to the human part and generate a melodic, harmonic and diverse machine part. Subjective evaluations on preferences show that the proposed algorithm generates music pieces of higher quality than the baseline method.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Traffic Prediction and Random Access Control Optimization: Learning and Non-learning based Approaches

Random access schemes in modern wireless communications are generally based on the framed-ALOHA (f-ALOHA), which can be optimized by flexibly organizing devices' transmission and re-transmission. However, this optimization is generally intractable due to the lack of information about complex traffic generation statistics and the occurrence of the random collision. In this article, we first summarize the general structure of access control optimization for different random access schemes, and then review the existing access control optimization based on Machine Learning (ML) and non-ML techniques. We demonstrate that the ML-based methods can better optimize the access control problem compared with non-ML based methods, due to their capability in solving high complexity long-term optimization problem and learning experiential knowledge from reality. To further improve the random access performance, we propose two-step learning optimizers for access control optimization, which individually execute the traffic prediction and the access control configuration. In detail, our traffic prediction method relies on online supervised learning adopting Recurrent Neural Networks (RNNs) that can accurately capture traffic statistics over consecutive frames, and the access control configuration can use either a non-ML based controller or a cooperatively trained Deep Reinforcement Learning (DRL) based controller depending on the complexity of different random access schemes. Numerical results show that the proposed two-step cooperative learning optimizer considerably outperforms the conventional Deep Q-Network (DQN) in terms of higher training efficiency and better access performance.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

High dimensional entanglement between a photon and a multiplexed atomic quantum memory

Multiplexed quantum memories and high-dimensional entanglement can improve the performance of quantum repeaters by promoting the entanglement generation rate and the quantum communication channel capacity. Here, we experimentally generate a high-dimensional entangled state between a photon and a collective spin wave excitation stored in the multiplexed atomic quantum memory. We verify the entanglement dimension by the quantum witness and the entanglement of formation. Then we use the high-dimensional entangled state to test the violation of the Bell-type inequality. Our work provides an effective method to generate multidimensional entanglement between the flying photonic pulses and the atomic quantum interface.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

Quantum Communication between Multiplexed Atomic Quantum Memories

The use of multiplexed atomic quantum memories (MAQM) can significantly enhance the efficiency to establish entanglement in a quantum network. In the previous experiments, individual elements of a quantum network, such as the generation, storage and transmission of quantum entanglement have been demonstrated separately. Here we report an experiment to show the compatibility of these basic operations. Specifically, we generate photon-atom entanglement in a $6\times 5$ MAQM, convert the spin wave to time-bin photonic excitation after a controllable storage time, and then store and retrieve the photon in a second MAQM for another controllable storage time. The preservation of quantum information in this process is verified by measuring the state fidelity. We also show that our scheme supports quantum systems with higher dimension than a qubit.

0 citations0 reviewsNo code linkedOpen access