arXiv subjects

Han Zheng

Publications and source records attributed to Han Zheng.

At least 37 records · Page 2Linked to original sources

arXiv2026

Learning-guided Prioritized Planning for Lifelong Multi-Agent Path Finding in Warehouse Automation

Lifelong Multi-Agent Path Finding (MAPF) is critical for modern warehouse automation, which requires multiple robots to continuously navigate conflict-free paths to optimize the overall system throughput. However, the complexity of warehouse environments and the long-term dynamics of lifelong MAPF often demand costly adaptations to classical search-based solvers. While machine learning methods have been explored, their superiority over search-based methods remains inconclusive. In this paper, we introduce Reinforcement Learning (RL) guided Rolling Horizon Prioritized Planning (RL-RH-PP), the first framework integrating RL with search-based planning for lifelong MAPF. Specifically, we leverage classical Prioritized Planning (PP) as a backbone for its simplicity and flexibility in integrating with a learning-based priority assignment policy. By formulating dynamic priority assignment as a Partially Observable Markov Decision Process (POMDP), RL-RH-PP exploits the sequential decision-making nature of lifelong planning while delegating complex spatial-temporal interactions among agents to reinforcement learning. An attention-based neural network autoregressively decodes priority orders on-the-fly, enabling efficient sequential single-agent planning by the PP planner. Evaluations in realistic warehouse simulations show that RL-RH-PP achieves the highest total throughput among baselines and generalizes effectively across agent densities, planning horizons, and warehouse layouts. Our interpretive analyses reveal that RL-RH-PP proactively prioritizes congested agents and strategically redirects agents from congestion, easing traffic flow and boosting throughput. These findings highlight the potential of learning-guided approaches to augment traditional heuristics in modern warehouse automation.

cs.AI↗

arXiv2026

SEAR: Schema-Based Evaluation and Routing for LLM Gateways

Evaluating production LLM responses and routing requests across providers in LLM gateways requires fine-grained quality signals and operationally grounded decisions. To address this gap, we present SEAR, a schema-based evaluation and routing system for multi-model, multi-provider LLM gateways. SEAR defines an extensible relational schema covering both LLM evaluation signals (context, intent, response characteristics, issue attribution, and quality scores) and gateway operational metrics (latency, cost, throughput), with cross-table consistency links across around one hundred typed, SQL-queryable columns. To populate the evaluation signals reliably, SEAR proposes self-contained signal instructions, in-schema reasoning, and multi-stage generation that produces database-ready structured outputs. Because signals are derived through LLM reasoning rather than shallow classifiers, SEAR captures complex request semantics, enables human-interpretable routing explanations, and unifies evaluation and routing in a single query layer. Across thousands of production sessions, SEAR achieves strong signal accuracy on human-labeled data and supports practical routing decisions, including large cost reductions with comparable quality.

cs.DB↗

arXiv2026

Efficient learning of logical noise from syndrome data

Characterizing errors in quantum circuits is essential for device calibration, yet detecting rare error events requires a large number of samples. This challenge is particularly severe in calibrating fault-tolerant, error-corrected circuits, where logical error probabilities are suppressed to higher order relative to physical noise and are therefore difficult to calibrate through direct logical measurements. Recently, Wagner et al. [PRL 130, 200601 (2023)] showed that, for phenomenological Pauli noise models, the logical channel can instead be inferred from syndrome measurement data generated during error correction. Here, we extend this framework to realistic circuit-level noise models. From a unified code-theoretic perspective and spacetime code formalism, we derive necessary and sufficient conditions for learning the logical channel from syndrome data alone and explicitly characterize the learnable degrees of freedom of circuit-level Pauli faults. Using Fourier analysis and compressed sensing, we develop efficient estimators with provable guarantees on sample complexity and computational cost. We further present an end-to-end protocol and demonstrate its performance on several syndrome-extraction circuits, achieving orders-of-magnitude sample-complexity savings over direct logical benchmarking. Our results establish syndrome-based learning as a practical approach to characterizing the logical channel in fault-tolerant quantum devices.

quant-ph↗

arXiv2026

OSNIP: Breaking the Privacy-Utility-Efficiency Trilemma in LLM Inference via Obfuscated Semantic Null Space

We propose Obfuscated Semantic Null space Injection for Privacy (OSNIP), a lightweight client-side encryption framework for privacy-preserving LLM inference. Generalizing the geometric intuition of linear kernels to the high-dimensional latent space of LLMs, we formally define the ``Obfuscated Semantic Null Space'', a high-dimensional regime that preserves semantic fidelity while enforcing near-orthogonality to the original embedding. By injecting perturbations that project the original embedding into this space, OSNIP ensures privacy without any post-processing. Furthermore, OSNIP employs a key-dependent stochastic mapping that synthesizes individualized perturbation trajectories unique to each user. Evaluations on 12 generative and classification benchmarks show that OSNIP achieves state-of-the-art performance, sharply reducing attack success rates while maintaining strong model utility under strict security constraints.

cs.LG↗

arXiv2026

Insights into the Physical Nature of Polar Ring Galaxies from H I Observations

Polar ring galaxies (PRGs) host an outer ring of gas and stars oriented nearly perpendicular to the main stellar body. They represent extreme examples of misaligned systems and provide valuable insight into galaxy interactions, gas accretion, and peculiar gas dynamics. We compile a complete sample of kinematically confirmed PRGs and collect their H I measurements. Combining literature data with new observations from FAST, we detect H I emission in 22 sources, identify one potential H I absorption feature, and find four non-detections among 40 confirmed PRGs. Compared to galaxies in the ALFALFA and xGASS surveys, PRGs predominantly occupy the green valley or quenched regimes but exhibit higher gas fractions than typical early-type galaxies, suggesting gas accretion. The H I profile asymmetry and shape for PRGs are not consistent with that of the ALFALFA sample with p<0.05. We examine their Tully-Fisher relation (TFR) and baryonic TFR (bTFR), linking the systems' rotation velocities to their masses. The extreme outliers in TFRs for the control sample tend to display single-peaked H I profiles. PRGs do not follow a tight TFR or bTFR if the H I resides primarily in the host galaxy. But the scatter decreases significantly if we assume the gas is mainly distributed in the polar ring. Spatially resolved H I observations are essential to disentangle the gas distribution and kinematics in PRGs, which are key to understanding their formation mechanisms.

astro-ph.GA↗

arXiv2025

Beyond Jailbreak: Unveiling Risks in LLM Applications Arising from Blurred Capability Boundaries

LLM applications (i.e., LLM apps) leverage the powerful capabilities of LLMs to provide users with customized services, revolutionizing traditional application development. While the increasing prevalence of LLM-powered applications provides users with unprecedented convenience, it also brings forth new security challenges. For such an emerging ecosystem, the security community lacks sufficient understanding of the LLM application ecosystem, especially regarding the capability boundaries of the applications themselves. In this paper, we systematically analyzed the new development paradigm and defined the concept of the LLM app capability space. We also uncovered potential new risks beyond jailbreak that arise from ambiguous capability boundaries in real-world scenarios, namely, capability downgrade and upgrade. To evaluate the impact of these risks, we designed and implemented an LLM app capability evaluation framework, LLMApp-Eval. First, we collected application metadata across 4 platforms and conducted a cross-platform ecosystem analysis. Then, we evaluated the risks for 199 popular applications among 4 platforms and 6 open-source LLMs. We identified that 178 (89.45%) potentially affected applications, which can perform tasks from more than 15 scenarios or be malicious. We even found 17 applications in our study that executed malicious tasks directly, without applying any adversarial rewriting. Furthermore, our experiments also reveal a positive correlation between the quality of prompt design and application robustness. We found that well-designed prompts enhance security, while poorly designed ones can facilitate abuse. We hope our work inspires the community to focus on the real-world risks of LLM applications and foster the development of a more robust LLM application ecosystem.

cs.CR↗

arXiv2025

Toward Super-polynomial Quantum Speedup of Equivariant Quantum Algorithms with SU($d$) Symmetry

We introduce a framework of the equivariant convolutional quantum algorithms which is tailored for a number of machine-learning tasks on physical systems with arbitrary SU$(d)$ symmetries. It allows us to enhance a natural model of quantum computation -- permutational quantum computing (PQC) -- and define a more powerful model: PQC+. While PQC was shown to be efficiently classically simulatable, we exhibit a problem which can be efficiently solved on PQC+ machine, whereas no classical polynomial time algorithm is known; thus providing evidence against PQC+ being classically simulatable. We further discuss practical quantum machine learning algorithms which can be carried out in the paradigm of PQC+.

quant-ph↗

arXiv2025

No-go theorems for logical gates on product quantum codes

Quantum error-correcting codes are essential to the implementation of fault-tolerant quantum computation. Homological products of classical codes offer a versatile framework for constructing quantum error-correcting codes with desirable properties, especially quantum low-density parity check (qLDPC) codes. Based on extensions of the Bravyi--König theorem that encompass codes without geometric locality, we establish a series of general no-go theorems for fault-tolerant logical gates supported by hypergraph product codes. Specifically, we show that non-Clifford logical gates cannot be implemented transversally on hypergraph product codes of all product dimensions, and that the dimensions impose various limitations on the accessible level of the Clifford hierarchy gates by constant-depth local circuits. We also discuss examples both with and without geometric locality which attain the Clifford hierarchy bounds. Our results reveal fundamental restrictions on logical gates originating from highly general algebraic structures, extending beyond existing knowledge only in geometrically local, finite logical qubits, transversal, or 2-dimensional product cases, and may guide the vital study of fault-tolerant quantum computation with qLDPC codes.

quant-ph↗

arXiv2025

SU(d)-Symmetric Random Unitaries: Quantum Scrambling, Error Correction, and Machine Learning

Quantum information processing in the presence of continuous symmetry is of wide importance and exhibits many novel physical and mathematical phenomena. SU(d) is a continuous group of particular interest since it represents a fundamental type of non-Abelian symmetry and also plays a vital role in quantum computation. Here, we explicate three particularly interesting applications of symmetric random unitaries in diverse contexts ranging from physics to quantum computing: information scrambling with non-Abelian conserved quantities, covariant quantum error correcting random codes, and geometric quantum machine learning. First, we show that, in the presence of SU(d) symmetry, the local conserved quantities would exhibit residual values even at $t \rightarrow \infty$ which decays as $Ω(1/n^{3/2})$ under local Pauli basis for qubits and $Ω(1/n^{(d+2)^2/2})$ under symmetric basis for general qudits with respect to the system size, in contrast to O(1/n) decay for U(1) case and the exponential decay for no-symmetry case in the sense of out-of-time ordered correlator. Second, we show that SU(d)-symmetric unitaries can be used to construct asymptotically optimal (in the sense of saturating the fundamental limits on the code error, or the approximate Eastin--Knill theorems) SU(d)-covariant codes that encode any constant number of logical qudits, extending [Kong & Liu; PRXQ 3, 020314 (2022)]. Finally, we derive an overpartameterization threshold via the quantum neural tangent kernel required for exponential convergence guarantee of generic ansatz for geometric quantum machine learning, which reveals that the number of parameters required scales only with the dimension of desired subspaces rather than the entire Hilbert space. Our work invites further research on quantum information with continuous symmetries, where the mathematical tools developed in this work are expected to be useful.

quant-ph↗

arXiv2025

Cuckoo Attack: Stealthy and Persistent Attacks Against AI-IDE

Modern AI-powered Integrated Development Environments (AI-IDEs) are increasingly defined by an Agent-centric architecture, where an LLM-powered Agent is deeply integrated to autonomously execute complex tasks. This tight integration, however, also introduces a new and critical attack surface. Attackers can exploit these components by injecting malicious instructions into untrusted external sources, effectively hijacking the Agent to perform harmful operations beyond the user's intention or awareness. This emerging threat has quickly attracted research attention, leading to various proposed attack vectors, such as hijacking Model Context Protocol (MCP) Servers to access private data. However, most existing approaches lack stealth and persistence, limiting their practical impact. We propose the Cuckoo Attack, a novel attack that achieves stealthy and persistent command execution by embedding malicious payloads into configuration files. These files, commonly used in AI-IDEs, execute system commands during routine operations, without displaying execution details to the user. Once configured, such files are rarely revisited unless an obvious runtime error occurs, creating a blind spot for attackers to exploit. We formalize our attack paradigm into two stages, including initial infection and persistence. Based on these stages, we analyze the practicality of the attack execution process and identify the relevant exploitation techniques. Furthermore, we analyze the impact of Cuckoo Attack, which can not only invade the developer's local computer but also achieve supply chain attacks through the spread of configuration files. We contribute seven actionable checkpoints for vendors to evaluate their product security. The critical need for these checks is demonstrated by our end-to-end Proof of Concept, which validated the proposed attack across nine mainstream Agent and AI-IDE pairs.

cs.CR↗

arXiv2025

MultiPark: Multimodal Parking Transformer with Next-Segment Prediction

Parking accurately and safely in highly constrained spaces remains a critical challenge. Unlike structured driving environments, parking requires executing complex maneuvers such as frequent gear shifts and steering saturation. Recent attempts to employ imitation learning (IL) for parking have achieved promising results. However, existing works ignore the multimodal nature of parking behavior in lane-free open space, failing to derive multiple plausible solutions under the same situation. Notably, IL-based methods encompass inherent causal confusion, so enabling a neural network to generalize across diverse parking scenarios is particularly difficult. To address these challenges, we propose MultiPark, an autoregressive transformer for multimodal parking. To handle paths filled with abrupt turning points, we introduce a data-efficient next-segment prediction paradigm, enabling spatial generalization and temporal extrapolation. Furthermore, we design learnable parking queries factorized into gear, longitudinal, and lateral components, parallelly decoding diverse parking behaviors. To mitigate causal confusion in IL, our method employs target-centric pose and ego-centric collision as outcome-oriented loss across all modalities beyond pure imitation loss. Evaluations on real-world datasets demonstrate that MultiPark achieves state-of-the-art performance across various scenarios. We deploy MultiPark on a production vehicle, further confirming our approach's robustness in real-world parking environments.

cs.RO↗

arXiv2025

Self-Aware Safety Augmentation: Leveraging Internal Semantic Understanding to Enhance Safety in Vision-Language Models

Large vision-language models (LVLMs) are vulnerable to harmful input compared to their language-only backbones. We investigated this vulnerability by exploring LVLMs internal dynamics, framing their inherent safety understanding in terms of three key capabilities. Specifically, we define these capabilities as safety perception, semantic understanding, and alignment for linguistic expression, and experimentally pinpointed their primary locations within the model architecture. The results indicate that safety perception often emerges before comprehensive semantic understanding, leading to the reduction in safety. Motivated by these findings, we propose \textbf{Self-Aware Safety Augmentation (SASA)}, a technique that projects informative semantic representations from intermediate layers onto earlier safety-oriented layers. This approach leverages the model's inherent semantic understanding to enhance safety recognition without fine-tuning. Then, we employ linear probing to articulate the model's internal semantic comprehension to detect the risk before the generation process. Extensive experiments on various datasets and tasks demonstrate that SASA significantly improves the safety of LVLMs, with minimal impact on the utility.

cs.AI↗

arXiv2025

CP-Router: An Uncertainty-Aware Router Between LLM and LRM

Recent advances in Large Reasoning Models (LRMs) have significantly improved long-chain reasoning capabilities over Large Language Models (LLMs). However, LRMs often produce unnecessarily lengthy outputs even for simple queries, leading to inefficiencies or even accuracy degradation compared to LLMs. To overcome this, we propose CP-Router, a training-free and model-agnostic routing framework that dynamically selects between an LLM and an LRM, demonstrated with multiple-choice question answering (MCQA) prompts. The routing decision is guided by the prediction uncertainty estimates derived via Conformal Prediction (CP), which provides rigorous coverage guarantees. To further refine the uncertainty differentiation across inputs, we introduce Full and Binary Entropy (FBE), a novel entropy-based criterion that adaptively selects the appropriate CP threshold. Experiments across diverse MCQA benchmarks, including mathematics, logical reasoning, and Chinese chemistry, demonstrate that CP-Router efficiently reduces token usage while maintaining or even improving accuracy compared to using LRM alone. We also extend CP-Router to diverse model pairings and open-ended QA, where it continues to demonstrate strong performance, validating its generality and robustness.

cs.CL↗

arXiv2025

Fixing 7,400 Bugs for 1$: Cheap Crash-Site Program Repair

The rapid advancement of bug-finding techniques has led to the discovery of more vulnerabilities than developers can reasonably fix, creating an urgent need for effective Automated Program Repair (APR) methods. However, the complexity of modern bugs often makes precise root cause analysis difficult and unreliable. To address this challenge, we propose crash-site repair to simplify the repair task while still mitigating the risk of exploitation. In addition, we introduce a template-guided patch generation approach that significantly reduces the token cost of Large Language Models (LLMs) while maintaining both efficiency and effectiveness. We implement our prototype system, WILLIAMT, and evaluate it against state-of-the-art APR tools. Our results show that, when combined with the top-performing agent CodeRover-S, WILLIAMT reduces token cost by 45.9% and increases the bug-fixing rate to 73.5% (+29.6%) on ARVO, a ground-truth open source software vulnerabilities benchmark. Furthermore, we demonstrate that WILLIAMT can function effectively even without access to frontier LLMs: even a local model running on a Mac M4 Mini achieves a reasonable repair rate. These findings highlight the broad applicability and scalability of WILLIAMT.

cs.SE↗

arXiv2025

Dark spin-cats as biased qubits

We present a biased atomic qubit, universally implementable across all atomic platforms, encoded as a `spin-cat' within ground state Zeeman levels. The key characteristic of our configuration is the coupling of the ground state spin manifold of size $F_g \gg 1$ to an excited Zeeman spin manifold of size $F_e = F_g - 1$ using light. This coupling results in eigenstates of the driven atom that include exactly two dark states in the ground state manifold, which are decoupled from light and immune to spontaneous emission from the excited states. These dark states constitute the `spin-cat', leading to the designation `dark spin-cat'. We demonstrate that under strong Rabi drive and for large $F_g$, the `dark spin-cat' is autonomously stabilized against common noise sources and encodes a qubit with significantly biased noise. Specifically, the bit-flip error rate decreases exponentially with $F_g$ relative to the dephasing rate. We provide an analysis of dark spin-cats, their robustness to noise, and discuss bias-preserving single qubit and entangling gates, exemplified on a Rydberg tweezer platform.

quant-ph↗

arXiv2025

Embodied Escaping: End-to-End Reinforcement Learning for Robot Navigation in Narrow Environment

Autonomous navigation is a fundamental task for robot vacuum cleaners in indoor environments. Since their core function is to clean entire areas, robots inevitably encounter dead zones in cluttered and narrow scenarios. Existing planning methods often fail to escape due to complex environmental constraints, high-dimensional search spaces, and high difficulty maneuvers. To address these challenges, this paper proposes an embodied escaping model that leverages reinforcement learning-based policy with an efficient action mask for dead zone escaping. To alleviate the issue of the sparse reward in training, we introduce a hybrid training policy that improves learning efficiency. In handling redundant and ineffective action options, we design a novel action representation to reshape the discrete action space with a uniform turning radius. Furthermore, we develop an action mask strategy to select valid action quickly, balancing precision and efficiency. In real-world experiments, our robot is equipped with a Lidar, IMU, and two-wheel encoders. Extensive quantitative and qualitative experiments across varying difficulty levels demonstrate that our robot can consistently escape from challenging dead zones. Moreover, our approach significantly outperforms compared path planning and reinforcement learning methods in terms of success rate and collision avoidance.

cs.RO↗

arXiv2025

D$^2$-DPM: Dual Denoising for Quantized Diffusion Probabilistic Models

Diffusion models have achieved cutting-edge performance in image generation. However, their lengthy denoising process and computationally intensive score estimation network impede their scalability in low-latency and resource-constrained scenarios. Post-training quantization (PTQ) compresses and accelerates diffusion models without retraining, but it inevitably introduces additional quantization noise, resulting in mean and variance deviations. In this work, we propose D2-DPM, a dual denoising mechanism aimed at precisely mitigating the adverse effects of quantization noise on the noise estimation network. Specifically, we first unravel the impact of quantization noise on the sampling equation into two components: the mean deviation and the variance deviation. The mean deviation alters the drift coefficient of the sampling equation, influencing the trajectory trend, while the variance deviation magnifies the diffusion coefficient, impacting the convergence of the sampling trajectory. The proposed D2-DPM is thus devised to denoise the quantization noise at each time step, and then denoise the noisy sample through the inverse diffusion iterations. Experimental results demonstrate that D2-DPM achieves superior generation quality, yielding a 1.42 lower FID than the full-precision model while achieving 3.99x compression and 11.67x bit-operation acceleration.

cs.CV↗

arXiv2025

Constrained Bandwidth Observation Sharing for Multi-Robot Navigation in Dynamic Environments via Intelligent Knapsack

Multi-robot navigation is increasingly crucial in various domains, including disaster response, autonomous vehicles, and warehouse and manufacturing automation. Robot teams often must operate in highly dynamic environments and under strict bandwidth constraints imposed by communication infrastructure, rendering effective observation sharing within the system a challenging problem. This paper presents a novel optimal communication scheme, Intelligent Knapsack (iKnap), for multi-robot navigation in dynamic environments under bandwidth constraints. We model multi-robot communication as belief propagation in a graph of inferential agents. We then formulate the combinatorial optimization for observation sharing as a 0/1 knapsack problem, where each potential pairwise communication between robots is assigned a decision-making utility to be weighed against its bandwidth cost, and the system has some cumulative bandwidth limit. We evaluate our approach in a simulated robotic warehouse with human workers using ROS2 and the Open Robotics Middleware Framework. Compared to state-of-the-art broadcast-based optimal communication schemes, iKnap yields significant improvements in navigation performance with respect to scenario complexity while maintaining a similar runtime. Furthermore, iKnap utilizes allocated bandwidth and observational resources more efficiently than existing approaches, especially in very low-resource and high-uncertainty settings. Based on these results, we claim that the proposed method enables more robust collaboration for multi-robot teams in real-world navigation problems.

cs.RO↗