Search arXiv⌕ Search

arXiv subjects

Qian Li

Publications and source records attributed to Qian Li.

At least 19 recordsLinked to original sources
arXiv2026

AquaMend: Minimal Re-probing and Conditional Rollback for Latent-Belief Failures in Embodied Agents

Physical changes or sensing errors can invalidate embodied agents' task-relevant beliefs. AquaMend compares re-probing, rollback, and supported continuation on a probe-belief-action graph under an expected-loss objective covering sensing, physical recovery, and uncorrected failures. A joint posterior guides a one-step policy with conditional detection-power screening. The per-belief three-way optimum requires independence, separability, and fully resolving probes; the general policy has no global optimality guarantee. Across 32 paired scenarios in a self-constructed simulation benchmark, AquaMend recovers in 28/32 cases and reduces mean complete loss by 21.6% versus restart. Its paired loss difference from decision-theoretic troubleshooting (DTT) is not statistically significant after Holm correction. Against the all-candidate ablation, online decision time decreases by 12.3% overall but increases by 3.4% in the uncovered late stage.

cs.RO↗
arXiv2026

Brain-Inspired Hierarchical Modularity for General Continual Learning

Continual learning, the ability to learn from sequential experience while retaining and adapting prior knowledge, is central to intelligent systems operating in changing environments. However, conventional continual learning is typically studied with offline task-wise training and clear task boundaries, leaving a substantial gap from general continual learning under online, uncertain, and evolving data streams. In this regime, intelligent systems must separate conflicting experience to reduce interference while integrating compatible experience to promote generalization. Inspired by the organization of the Drosophila learning and memory system, we identify a hierarchical modular principle that coordinates both functions through expert specialization and ensemble integration. We instantiate this principle as lightweight modular adaptation of pretrained foundation models, combining brain-inspired random expansion for expert routing and diversified modular integration across spatial and temporal scales. Across visual recognition, vision-language understanding, ego-exo video understanding, and embodied vision-language-action learning, our method consistently improves learning under online and uncertain data streams, with gains exceeding 50 percentage points over replay-free alternatives in embodied manipulation. These findings support hierarchical modularity as a biologically grounded path for learning from dynamic experience.

cs.LG↗
arXiv2026

Entanglement free Metrology Exploiting Multimode Hong Ou Mandel Sensor Advantage

The Hong-Ou-Mandel (HOM) interference in the multimode frequency domain has been explored for precision metrology, with several experimental demonstrations exploiting its robustness against dispersion and phase noise, as well as its large dynamic range and compatibility with fragile samples. Conventional multimode HOM metrology exploits frequency-entangled states, which naturally satisfy bosonic exchange symmetry under any centered symmetric joint spectral distribution, to provide these advantages. However, these entangled states are typically generated via spontaneous parametric down-conversion (SPDC), requiring strong pump lasers that hinder practical implementation. In this paper, we employ frequency product states, which do not possess entanglement or path-mode exchange symmetry, as the probe state and post-select measurement outcomes exhibiting frequency anti-correlation. Our results demonstrate that these advantages,peak narrowing, dispersion cancellation, phase-noise immunity, a large dynamic range, and compatibility with fragile samples, arise neither from entanglement nor from bosonic exchange symmetry, but rather from spectral anti-correlation. We further show that entanglement is not the source of the measurement precision: the entanglement-free approach attains the same quantum Fisher information as the entangled-state scheme, indicating that the fundamental precision limit does not originate from entanglement.

quant-ph↗
arXiv2026

REVOLVE: An Automated Closed-Loop Framework for Evolving Robot Manipulation with Minimal Human Intervention

Recent advances in data-driven robot manipulation policies have substantially improved task execution and generalization. However, real-world deployment still relies heavily on humans for failure assessment, correction, and environment reset, while models often fail to continually learn from failures and corrective experience. We present REVOLVE (Robot Evolving via Orchestrated Loops, Verification, and Experience), an automated closed-loop framework for evolving robot manipulation with minimal human intervention. Built on a unified software platform, REVOLVE integrates data collection, policy training and deployment, failure recovery, and continual learning into a single closed-loop workflow. Its Automated Reset and Correction (ARC) architecture automatically resets the environment and intervenes to correct policy failures. Dual-Loop Evolution (DLE) continually improves the manipulation policy and agent by feeding real-world interaction and failure--correction data back into policy learning and using an external mismatch memory to refine agent judgments. Experiments across four real-world manipulation tasks show that, after five iterations, REVOLVE improves average policy success rate by 18.5% and agent judgment accuracy by 8.5%, while reducing human effort in data collection and deployment testing by 94.4% and 95.1%, respectively. These results demonstrate that REVOLVE transforms real-world deployment into a closed-loop learning process that continually accumulates and uses execution experience, enabling continual evolution of both the policy and supervisory model with substantially less human intervention.

cs.RO↗
arXiv2026

Global Stability of 3D Compressible Non-Resistive MHD: Hidden Damping and Rational Background Fields

We prove the global well-posedness and nonlinear stability of classical solutions to the three-dimensional compressible viscous, non-resistive MHD system on $\mathbb{T}^3$ near the equilibrium $(1,\mathbf{0},\mathbf{e}_3)$, for small $x_3$-symmetric perturbations and without any Diophantine condition on the background magnetic field. The central obstruction is the $x_3$-independent sector, in which the density and the magnetic field possess no dissipation, no damping, and no decay. We overcome it by exhibiting a hidden wave structure for the exact total pressure $\mathcal{D}=P(1+a)+B_3+\frac12|\mathbf{B}|^2$, the averaged pair $(\overline{\mathbf{u}},\overline{\mathcal{D}})$ obeys a closed system, and $\overline{\mathcal{D}}$ satisfies a strongly damped wave equation whose principal wave part propagates at the fast magnetosonic speed and whose non-parabolic branch damps at a rate that involves no gain of derivatives. This hidden damping substitutes for the missing magnetic dissipation, and simultaneously absorbs the magnetic pressure $\frac12\nabla|\mathbf{B}|^2$, the main obstruction created by compressibility. Together with a damped wave structure for the oscillatory sector and space-time weighted energy functionals with shifted time weights, this yields global existence, uniform stability, and explicit polynomial decay rates.

math.AP↗
arXiv2026

AMA: Adaptive Memory via Multi-Agent Collaboration

The rapid evolution of Large Language Model (LLM) agents has necessitated robust memory systems to support cohesive long-term interaction and complex reasoning. Benefiting from the strong capabilities of LLMs, recent research focus has shifted from simple context extension to the development of dedicated agentic memory systems. However, existing approaches typically rely on rigid retrieval granularity, accumulation-heavy maintenance strategies, and coarse-grained update mechanisms. These design choices create a persistent mismatch between stored information and task-specific reasoning demands, while leading to the unchecked accumulation of logical inconsistencies over time. To address these challenges, we propose Adaptive Memory via Multi-Agent Collaboration (AMA), a novel framework that leverages coordinated agents to manage memory across multiple granularities. AMA employs a hierarchical memory design that dynamically aligns retrieval granularity with task complexity. Specifically, the Constructor and Retriever jointly enable multi-granularity memory construction and adaptive query routing. The Judge verifies the relevance and consistency of retrieved content, triggering iterative retrieval when evidence is insufficient or invoking the Refresher upon detecting logical conflicts. The Refresher then enforces memory consistency by performing targeted updates or removing outdated entries. Extensive experiments on challenging long-context benchmarks show that AMA significantly outperforms state-of-the-art baselines while reducing token consumption by approximately 80% compared to full-context methods, demonstrating its effectiveness in maintaining retrieval precision and long-term memory consistency.

cs.AI↗
arXiv2026

On the Expressive Power and Limitations of Multi-Layer SSMs

We study how depth, finite precision, state dimension, and chain-of-thought (CoT) affect the expressive power of multi-layer state-space models (SSMs). For the explicit-table $K$-function-composition problem, a canonical benchmark for sequential information propagation, we prove that any $L$-layer SSM solving $(L+3)$-function composition must satisfy $d^2p=Ω(N/L^3)$, where $d$ is the state dimension and $p$ is the per-scalar precision. Conversely, $K$-function composition is solved exactly by a $(K+1)$-layer generalized SSM with $d=1$ and $p=Θ(\log N)$. This gives a worst-case depth hierarchy for this formal problem family. We then distinguish post-input reasoning, in which all thought tokens are generated after the input, from input-interleaved reasoning, in which thought tokens may be inserted while the input stream is being read. Post-input reasoning does not circumvent our communication-based lower-bound pipeline, whereas input-interleaved reasoning admits bidirectional simulations with general deterministic one-pass streaming algorithms at the granularity of persistent memory. Finally, width and precision are not interchangeable under exact step-preserving simulation in the base affine-state model, but become interchangeable through the streaming-memory characterization once input-interleaved reasoning is allowed.

cs.LG↗
arXiv2026

Hindsight Memory-PRM: Supervising Memory Management with Auditable Hindsight Credit

Memory operations of long-horizon LLM agents are hard to supervise: an operation's value is unobservable when it is taken. But they are special -- they leave machine-readable evidence in the trajectory: retrieval hits and answer-time citations. Hindsight Memory-PRM exploits this audit trail twice: offline to train an operation-conditioned memory-utility critic, and online, where retrievals, citations, and one controlled deletion-and-reanswer per probe settle an intervention-calibrated entry-level presence credit, propagated along version chains as an action-level proxy reward -- no per-operation human labels, no Monte-Carlo replay of continuations. On held-out LoCoMo a local 8B policy reaches 77.5% under a fixed shared reader, surpassing its API teacher (65.1%) and all reproduced external systems, at one eighth the context of Mem0's official operating point; on LongMemEval, 79.0%. Ablations attribute the gain to causal calibration rather than signal density, and the policy converges to a multi-version memory organization whose gains no tested open-loop baseline reproduces.

cs.CL↗
arXiv2026

A Target-Centric Survey of Quantization-Aware Training

The rapid development of LLMs incurs prohibitive memory footprints and intensive computational demands. Quantization-Aware Training (QAT) techniques have emerged as a promising solution to address these challenges by explicitly simulating quantization effects during model training, yielding low-bit models that achieve accuracy comparable to their full-precision counterparts. In this work, we provide a target-centric survey of QAT, aimed at clarifying both its theoretical foundations and its evolving implementation landscape. We systematically review existing QAT methods through a target-centric taxonomy and synthesize cross-target differences in error characteristics, numerical formats, and strategy transferability. We further summarize QAT evaluation paradigms and discuss challenges in optimization and deployment, outlining potential directions for future research.

cs.LG↗
arXiv2026

Test-time Reinforcement Learning for Anomalous Video Understanding

Anomalous video understanding aims to identify abnormal events in videos and interpret their semantic meanings beyond simple anomaly detection. Recent video large language models (Video-LLMs) have demonstrated promising zero-shot capabilities for this task, yet their performance remains limited due to insufficient adaptation to diverse anomaly patterns and evolving environments. Test-time reinforcement learning offers a promising solution by enabling models to improve through self-generated feedback signals without requiring additional human annotations. However, applying it to anomalous video understanding remains challenging due to three issues: (1) generated pseudo-labels can be unreliable when consensus is weak; (2) binary reward designs fail to capture uncertainty in model generations, resulting in ineffective optimization signals; and (3) unanimous rollout groups receive identical rewards, causing group-relative advantages to collapse and eliminating effective policy-gradient signals. To address these challenges, we present a novel test-time reinforcement learning framework for anomalous video understanding by introducing dual-query consistency filtering, an entropy-aware consensus reward, and a virtual negative anchor mechanism. The framework retains reliable samples through consistency across semantically equivalent queries, combines answer agreement with generation uncertainty for reward estimation, and introduces a virtual negative anchor to create reward variation in unanimous rollout groups, thereby preserving effective group-relative optimization signals. Experiments on VAU-Bench show that our method outperforms the compared frozen and supervised baselines. The gains are most pronounced on the ECVA subset of VAU-Bench with thinking, where accuracy improves from 75.81% to 90.00% relative to the frozen backbone.

cs.CV↗
arXiv2026

Radiation properties and images of loop quantum Reissner-Nordström black hole with a thin accretion disk

We investigate the characteristics of circular motion of charged particles around loop quantum Reissner-Nordström black hole (LQRNBH) and the radiation properties and observational appearance of a thin accretion disk around it. By calculating the shadow radius and utilizing observational data from M87* and Sgr A*, we derive constraints on the quantum parameter $ζ$ and charge parameter $Q$. In order to support a accretion disk around the black hole, the charge parameter $Q$ is restricted to be tiny ($Q\lesssim10^{-21}$). The circular motion of charged particles around LQRNBH and the influence of the model parameters on the circular motion are studied. It is found that although the direct charge correction to the geometry is negligible, the electromagnetic coupling $κ$ substantially impacts the circular motion and the radiative properties. Then, by considering a thin accretion disk model, the local radiation fluxes, redshift factors and the observed fluxes are studied. With the ray-tracing method, the isoradial curves, direct images of redshift factors and the observed fluxes are illustrated. The influence of the model parameters on these quantities are systematically performed. It is found that the influence of the quantum parameter $ζ$ is generally very small but whether its a positive or negative effect depends on the sign of $κ$, i.e. attractive or repulsive electromagnetic force. Although the charge of the black hole is tiny in our study, the electromagnetic interaction generally has significant influence on the physical quantities.

gr-qc↗
arXiv2026

Polarization Vortices in a Ferromagnetic Metal via Twistronics

Recent advances in moire engineering provide new pathways for manipulating lattice distortions and electronic properties in low-dimensional materials. Here, we demonstrate that twisted stacking can induce dipolar vortices in metallic SrRuO3 membranes, despite the presence of free charges that would normally screen depolarizing fields and dipole-dipole interactions. These polarization vortices are correlated with moire-periodic flexoelectricity induced by shear strain gradients, and exhibit a pronounced dependence on the twist angle. In addition, multiferroic behavior emerges below the ferromagnetic Curie temperature of the films, whereby polarization and ferromagnetism coexist and compete, showing opposite twist-angle dependencies of their respective magnitudes. Density functional theory calculations provide insights into the microscopic origin of these observations. Our findings extend the scope of polarization topology design beyond dielectric materials and into metals.

cond-mat.mtrl-sci↗
arXiv2026

KnowHal: A Knowledge-Driven Benchmark for Comprehensive Multimodal Hallucination Evaluation

Hallucination remains a critical challenge for developing trustworthy Multimodal Large Language Models (MLLMs). While existing benchmarks mainly focus on entity, attribute, and relation hallucinations, knowledge-related failures are often investigated separately, lacking a unified evaluation framework across different hallucination dimensions. To overcome this, we propose \textbf{KnowHal}, a benchmark that explicitly incorporates knowledge hallucination into multimodal hallucination evaluation spanning four dimensions: entity, attribute, relation, and knowledge. KnowHal constructs paired positive and negative questions over shared images and entities, enabling controlled comparisons among perceptual errors, knowledge-related errors, and false-premise acceptance. The benchmark contains 1,800 samples across 10 domains and 50 categories, constructed through a semi-automated pipeline combining LLM assistance, CLIP-based filtering, and human verification. We evaluate 14 representative MLLMs on KnowHal and conduct extensive analyses. Results show that the knowledge dimension consistently presents the greatest challenge for nearly all evaluated models, while most models exhibit substantial performance degradation on negative questions, revealing limited robustness to false premises. By unifying four hallucination dimensions with paired question design, KnowHal addresses an important gap in existing evaluation frameworks and enables a more comprehensive assessment of hallucinations in MLLMs.

cs.AI↗
arXiv2026

Impossibility of Perfectly Complete Many-Round Key Agreement in the QROM

This paper proves that it is impossible to construct perfectly complete quantum key agreement protocols (QKA) from quantumly secure one-way functions (OWFs) in a black-box manner. Specifically, consider any protocol in which Alice and Bob exchange only classical messages, make at most $q_{\mathsf{A}}$ and $q_{\mathsf{B}}$ quantum queries, respectively, to a Boolean-valued random oracle, and agree on a shared key with certainty. This paper shows that there exists an eavesdropper, given the classical messages, that can recover the shared key with certainty using $O((q_{\mathsf{A}}+q_{\mathsf{B}})^5)$ classical oracle queries. The bound is independent of the number of rounds, transcript length, key length, and oracle-domain size. Previous results only applies to two-round key agreement (Li et al. CRYPTO 26) or relies on unproven conjectures (Austrin et al. CRYPTO 22). GPT-5.6 Sol Ultra found this proof in a one-shot conversation and drafted a preliminary version of this paper. The authors are fully responsible for the correctness, writing and discussions of this paper.

quant-ph↗
arXiv2026

BRiG-AFA: Bellman Risk-to-Go Learning for Non-Myopic Active Feature Acquisition

Active feature acquisition (AFA) asks which unobserved feature to measure next for each test instance under a budget. Greedy rules are easy to train but can overlook context features whose value is realized only through later acquisitions, while reinforcement-learning and generative approaches introduce difficult optimization or conditional-density estimation. We introduce \method, a deployable, supervised alternative that learns a separate candidate-conditioned risk-to-go function for every remaining budget. Starting from the one-step terminal classification risk, the functions are fitted backward with Bellman targets; inference greedily minimizes the learned terminal risk using only observed values, the mask, candidate identity, and remaining budget. A controlled non-myopic benchmark shows the expected mechanism: at budgets two and three, \method improves accuracy over its one-step ablation by $4.84\pm2.17$ and $4.39\pm1.10$ percentage points (mean $\pm$ standard error over five seeds). On Fashion-MNIST with 20 candidate pixels, it improves accuracy at every nontrivial reported budget on average, including $10.20\pm0.74$ points at four acquisitions; its mean paired gain across budgets $\{2,4,8,12,16\}$ is $3.50\pm0.37$ points. A three-seed MiniBooNE study is mixed at small budgets but positive at 8 and 16 acquisitions, identifying a current boundary rather than supporting a universal claim. These results establish a reproducible mechanism-level case for direct Bellman risk regression and delimit the experiments still needed for state-of-the-art comparison.

cs.LG↗
arXiv2026

The User Asks, Platforms Compete: How Agentic Recommendation Markets Take Shape

Online recommendation has traditionally taken place after a user enters a platform, which determines the candidate pool and the ranking shown to the user. LLM-based user agents enable a different recommendation process: a user specifies a need before choosing a platform, leaving platforms to compete for the user's attention, which we refer to as an agentic recommendation market. In our controlled LLM-based experiments across three product domains, we find this new setting of recommendation creates a tension between access and attention. Compared with traditional platform-centric recommendation, user-centric recommendation greatly expands the opportunity for relevant items to enter comparison; yet broader participation does not translate directly into effective exposure. Competition directly triggers platforms' strategic play: selectively positive explanations occupy 73--78% of first-ranked positions. When the user agent relates platforms' actions to subsequent user feedback, this share falls to 36--41%, while the chance of a user purchasing the relevant item increases. A user agent is therefore more than a ranker over a larger pool of candidates: its querying, ranking, and feedback mechanism governing who can compete, how scarce attention is allocated, and how earlier outcomes shape the evaluation of platforms directly affect user utility. Designing agentic recommendation therefore requires treating access, attention, and accountability as a joint mechanism design problem.

cs.AI↗
arXiv2026

A Cybersecurity MLPS Large Language Model with Multi-Path Retrieval Fusion

The Multi-Level Protection Scheme (MLPS) is a foundational system in China's cybersecurity governance framework. Therefore, accurate analysis and understanding of MLPS requirements are essential. At present, MLPS analysis still relies mainly on manual interpretation of standards and rule-based tools. This makes it hard to provide stable and consistent compliance analysis in complex application scenarios. The rise of large language models has created new opportunities for making MLPS work more intelligent. However, in standards-intensive and security-sensitive scenarios, general-purpose large language models often cannot ensure controllable reasoning or complete understanding of rules. This paper proposes a large language model framework for MLPS that integrates multiple retrieval strategies. It combines hierarchical retrieval, tree-based retrieval, and tokenization-based matching retrieval. This design helps maintain retrieval coverage while reducing the interference of irrelevant context in the reasoning process. To address the requirements of MLPS question answering for clause accuracy, conclusion traceability, and practical deployability, this paper adopts a evaluation method based on multi-dimensional weighted scoring to quantitatively assess model responses. In comparative experiments on ten typical questions, the proposed domain-specific large language model for MLPS achieved higher overall scores.

cs.CR↗
arXiv2026

Global Existence and Pathwise Uniqueness for a Stochastic Parabolic-Parabolic Keller-Segel System

In this paper, we study a stochastic parabolic-parabolic Keller-Segel system driven by nonlocal, nonlinear multiplicative noise in a two-dimensional bounded domain. Under suitable assumptions, we establish global existence and pathwise uniqueness of a strong solution for arbitrary initial data, without imposing any smallness conditions. This sharply contrasts with the deterministic two-dimensional Keller-Segel system, which typically requires smallness assumptions on initial mass. The main analytical challenges stem from the fully parabolic coupling, combined with a lack of coercivity and global Lipschitz continuity in both the chemotactic drift and noise terms. To overcome this, we introduce a tailored truncated system to establish local existence via Banach's fixed point theorem. Using this local existence and pathwise uniqueness, we construct a maximal local strong solution. Finally, by introducing a specialized Lyapunov functional, we derive uniform estimates to extend this solution globally.

math.AP↗