Searcharxiv⌕ Search

arXiv subjects

Jun Wang

Publications and source records attributed to Jun Wang.

At least 91 records · Page 5Linked to original sources
arXiv2026

UniMem: Complementary Episodic-to-Parametric Memory for Boundary-Agnostic Task Streams

Memory is essential for LLM agents to accumulate task experience and reuse task-specific execution strategies. However, real-world deployment over boundary-agnostic and evolving task streams exposes a fundamental stability-plasticity dilemma. External retrieval-based memory can rapidly absorb new evidence, but it often fails to internalize recurring execution patterns and incurs inference-time retrieval overhead. Parametric memory enables stable and efficient execution once learned, but typically relies on explicit task boundaries and fixed parameter budgets. Inspired by the human brain, which balances plasticity and stability through complementary episodic storage and gradual consolidation, we propose UniMem, a self-routing framework for autonomous memory management. UniMem uses learnable routing tokens as memory controllers, enabling adaptive coordination between complementary memory pathways: novel or sparse tasks are retained in an episodic buffer for retrieval-augmented execution, while recurring and reliable patterns are consolidated into expandable parametric memory. By decoupling task identification from task execution with routing tokens and parametric memory blocks, UniMem expands memory on demand without task labels during deployment or uncontrolled parameter growth. Experiments on long-horizon streaming task sequences show that UniMem consistently outperforms baselines while maintaining execution fidelity, achieving an average gain of 4.0 EM points across three backbone models.

cs.CL↗
arXiv2026

COCO-OLAC: A Benchmark for Occluded Panoptic Segmentation and Image Understanding

To help address the occlusion problem in panoptic segmentation and image understanding, this paper proposes a new large-scale dataset named COCO-OLAC (COCO Occlusion Labels for All Computer Vision Tasks), which is derived from the COCO dataset by manually labelling images into three perceived occlusion levels. Using COCO-OLAC, we systematically assess and quantify the impact of occlusion on panoptic segmentation on samples having different levels of occlusion. Comparative experiments with SOTA panoptic models demonstrate that the presence of occlusion significantly affects performance, with higher occlusion levels resulting in notably poorer performance. Additionally, we propose a straightforward yet effective method as an initial attempt to leverage the occlusion annotation using contrastive learning to render a model that learns a more robust representation capturing different severities of occlusion. Experimental results demonstrate that the proposed approach boosts the performance of the baseline model and achieves SOTA performance on the proposed COCO-OLAC dataset.

cs.CV↗
arXiv2026

LinearARD: Linear-Memory Attention Distillation for RoPE Restoration

The extension of context windows in Large Language Models is typically facilitated by scaling positional encodings followed by lightweight Continual Pre-Training (CPT). While effective for processing long sequences, this paradigm often disrupts original model capabilities, leading to performance degradation on standard short-text benchmarks. We propose LinearARD, a self-distillation method that restores Rotary Position Embeddings (RoPE)-scaled students through attention-structure consistency with a frozen native-RoPE teacher. Rather than matching opaque hidden states, LinearARD aligns the row-wise distributions of dense $Q/Q$, $K/K$, and $V/V$ self-relation matrices to directly supervise attention dynamics. To overcome the quadratic memory bottleneck of $n \times n$ relation maps, we introduce a linear-memory kernel. This kernel leverages per-token log-sum-exp statistics and fuses logit recomputation into the backward pass to compute exact Kullback-Leibler divergence and gradients. On LLaMA2-7B extended from 4K to 32K, LinearARD recovers 98.3\% of the short-text performance of state-of-the-art baselines while surpassing them on long-context benchmarks. Notably, our method achieves these results using only \textbf{4.25M} training tokens compared to the \textbf{256M} tokens required by LongReD and CPT. Our code is available at https://github.com/gracefulning/LinearARD.

cs.CL↗
arXiv2026

Geometry-Induced Vacuum Polarization and Mode Shifts in Maxwell-Klein-Gordon Theory

Geometric confinement is known to modify single-particle dynamics through effective potentials, yet its imprint on the interacting quantum vacuum remains largely unexplored. In this work, we investigate the Maxwell--Klein--Gordon system constrained to curved surfaces and demonstrate that the geometric potential $Σ_{\mathrm{geom}}(\mathbf{r})$ acts as a local renormalization environment. We show that extrinsic curvature modifies the scalar loop spectrum, entering the vacuum polarization as a position-dependent mass correction $M^2(\mathbf{r}) \to m^2 + Σ_{\mathrm{geom}}(\mathbf{r})$. This induces a finite, gauge-invariant ``geometry-induced running'' of the electromagnetic response. In the long-wavelength regime ($|{\bf Q}|R \ll 1$), we derive a closed-form expression for the relative frequency shift $Δω/ω$, governed by the overlap between the electric energy density and the geometric potential. Applying this formalism to Gaussian bumps, cylindrical shells, and tori, we identify distinct spectral signatures that distinguish these quantum loop corrections from classical geometric optics. Our results suggest that spatial curvature can serve as a tunable knob for ``vacuum engineering,'' offering measurable shifts in high-$Q$ cavities and plasmonic systems.

physics.optics↗
arXiv2026

Occlusion-Aware Panoptic Segmentation with Joint Position Embedding and Occlusion-Level Attention

Panoptic segmentation in complex scenes remains challenging because of occlusions, yet modern approaches often neglect occlusion modelling. In this paper, we propose Position Embedding Modulation with Occlusion Level Attention (PEMOLA), a novel occlusion-aware module that can be seamlessly integrated into transformer-based panoptic segmentation. To obtain occlusion cues, we train an occlusion classifier on the COCO-OLAC dataset. The classifier derives the occlusion-level attention, which serves as spatial guidance, while the occlusion labels are encoded into a learnable embedding to produce channel-wise weights. Through joint modulation, PEMOLA elegantly introduces the occlusion priors into the position embedding, thereby improving the occlusion modelling. We further annotate the Cityscapes dataset with occlusion levels, termed Cityscapes Occlusion Labels for All Computer Vision Tasks (Cityscapes-OLAC), following the same labelling protocol as COCO-OLAC, to evaluate the cross-dataset generalisation ability of PEMOLA. Extensive experiments on COCO-OLAC and Cityscapes-OLAC demonstrate that PEMOLA consistently improves panoptic segmentation quality while introducing minimal computational overhead. These results highlight the importance of occlusion modelling, where incorporating occlusion-level attention helps deliver robust panoptic segmentation under occlusion. Code and dataset are available at https://github.com/wenbo-wei/PEMOLA.

cs.CV↗
arXiv2026

Does Super-Resolution Preserve Defect Evidence? A Low-False-Call Benchmark for Semiconductor Inspection

Super-resolution can make inspection images appear sharper without preserving the evidence needed to detect a defect. We study this failure mode with a benchmark that separates reconstruction from detection and evaluates both at a predeclared low false-positive rate. Ten end-to-end repetitions combine independently generated line/space and contact-hole images with model training, calibration, clean controls, weak defects, and a held-out defect morphology. Every reconstruction is scored by the same local residual detector, while direct and jointly trained detectors form a separate comparison track. Reconstruction fidelity and inspection utility diverge: the two learned reconstruction models attain the highest structural similarity yet detect fewer defect pixels than bicubic interpolation in every paired repetition. A direct DeepLabV3 detector reaches $0.1984\pm0.0385$ recall at $0.000174\pm0.000084$ false-positive rate and satisfies the held-out feasibility criterion in all ten repetitions. An illustrative joint model, DPU-WaferSR, passes independent clean calibration but exceeds the held-out limit in all ten repetitions, demonstrating that calibration success does not guarantee transfer. Weak-defect recall remains near zero for every feasible method. Applying the unchanged policies to 4,591 public Carinthia-S masks further reveals large method-dependent shifts on real SEM texture. These results support a simple conclusion: super-resolution for inspection should be judged by preserved task evidence and operating-point transfer, not reconstruction quality alone.

cs.CV↗
arXiv2026

On two-dimensional steady compactly supported Euler flows with constant vorticity

In this paper, we study the two-dimensional steady compactly supported incompressible Euler equations with free boundaries. We consider flows with constant vorticity that are perturbations of annular equilibria, in contrast to the laminar flows that predominate in the existing literature on steady water waves. More precisely, we analyze three distinct classes of steady Euler flows with compact support, which correspond, respectively, to partially overdetermined, two-phase overdetermined, and (fully) overdetermined elliptic free-boundary problems. Our main contributions are threefold. For each class, we first prove a flexibility result-the existence of nontrivial admissible domains-by combining shape derivatives with local bifurcation theory. Second, we establish the corresponding rigidity results. Third, we apply the implicit function theorem to show that the standard annular flows are stable under small perturbations of the Neumann boundary condition. These results provide new perspectives on the theory of overdetermined elliptic problems.

math.AP↗
arXiv2026

Pinching-Antenna Systems (PASS)-Based User-Side Navigation: An Anchor-Line-based Approach

Pinching-antenna systems (PASS) are capable of dynamically reconfiguring wireless channels by flexibly repositioning pinching antennas (PAs) along the waveguides to establish short-range line-of-sight links. In this paper, a user-side navigation framework for PASS is proposed, where mobile users determine their own positions using only downlink broadcast signals without any prior knowledge of the PA positions. First, a Lambert W function-based PA positioning and pseudorange estimation (LWF-PAP) algorithm is developed, in which the closed-form expressions for both the PA positions along the waveguide and the PA-user pseudoranges are derived. Second, a weighted least squares-based PASS navigation (WLS-PAN) algorithm is formulated, where the nonlinear PASS-based navigation equations are transformed into a closed-form linear system, and the optimal weight matrix is derived, achieving minimum-variance unbiased estimation. Third, the PA-derived position dilution of precision (PA-PDOP) metric is further defined to characterize the theoretical accuracy bound. Simulation results demonstrate that centimeter-level positioning accuracy is achieved for both PAs and users within the breakpoint distance. It is also shown that uniform PA deployment and a moderate increase in the number of PAs effectively improve navigation accuracy, thereby validating the effectiveness and robustness of the proposed framework for distributed real-time user-side self-navigation.

eess.SP↗
arXiv2026

On the Performance of Pinching-Antenna Systems (PASS) Under Dynamic Channels with Blockages

The performance of pinching-antenna systems (PASS) is fundamentally affected by line-of-sight (LoS) blockage in practical environments. In this paper, PASS is investigated under realistic, obstacle-induced blockage by jointly considering the LoS and non-LoS (NLoS) components, rather than relying on a LoS channel or a probabilistic blockage model. A geometry-aware blockage model is adopted, where a blockage region on the waveguide is defined according to the actual locations and geometric features of obstacles, such that a pinching-antenna (PA) located within the blockage region is unable to establish a LoS link to the user equipment (UE). The channel models of PASS are developed by jointly accounting for in-waveguide attenuation and spatial propagation loss. To quantify the impact of channel factors on PASS performance, a single-PA single-UE scenario is studied under Rayleigh and Rician fading channels. Closed-form expressions for the outage probability are derived for both cases. For the ergodic rate, a closed-form expression is obtained in the Rayleigh case, while a complete analytical expression and an approximate closed-form expression are derived in the Rician case. Analytical expressions are derived for the endpoints of the blockage region, and the deployment criteria of optimal PA are provided. Simulation results validate the analysis and reveal that: i) NLoS scattering has a twofold effect on PASS performance, potentially degrading the outage performance while improving the rate performance under Rician fading; ii) Sufficiently strong NLoS scattering can still sustain communication in the presence of LoS blockage; iii) The optimal PA position is jointly determined by the environment geometry and the interplay between spatial propagation loss and in-waveguide attenuation.

eess.SP↗
arXiv2026

On the Blockage Effect in Pinching-Antenna Systems (PASS)

Pinching-antenna systems (PASS) offer considerable potential for wireless communications due to their unique ability to dynamically reconfigure radiation positions along a waveguide. However, the performance of PASS remains a critical challenge in the presence of random line-of-sight (LoS) blockage, leading to significant attenuation and even communication outages. In this paper, the performance of PASS in the presence of LoS blockage is investigated from the perspective of stochastic geometry. Obstacles are modeled through a homogeneous Poisson point process (PPP), where the geometric dimensions, numbers, and positions are treated as random variables. To conduct a concrete characterization of LoS blockage, the random-height-and-random-radius (RHRR) blockage model and the deterministic-height-and-deterministic-radius (DHDR) blockage model are proposed. In particular, closed-form analytical and asymptotic expressions for the outage probability are obtained, along with analytical and approximate expressions for the ergodic rate. Our simulation results reveal that denser obstacle environments or statistically larger obstacles substantially increase the probability of LoS blockage and degrade the system performance. Moreover, owing to its ability to dynamically reposition PAs, PASS can consistently outperform conventional antenna systems in the presence of LoS blockage.

eess.SP↗
arXiv2026

DeepTravel: An End-to-End Agentic Reinforcement Learning Framework for Autonomous Travel Planning Agents

Travel planning (TP) agent has recently worked as an emerging building block to interact with external tools/resources for travel itinerary generation, ensuring an enjoyable user experience. Despite its benefits, existing studies rely on hand-craft prompt and fixed agent workflow, hindering more flexible and autonomous TP agents. This paper proposes DeepTravel, an end-to-end agentic reinforcement learning framework for building an autonomous travel planning agent, capable of autonomously planning, executing tools, and reflecting on tool responses to explore, verify, and refine intermediate actions in multi-step reasoning. To achieve this, we first construct a robust travel sandbox by caching transportation, accommodation and POI data, facilitating TP agent training without being constrained by real-world APIs limitations (e.g., inconsistent outputs). Moreover, we develop a hierarchical reward modeling system, where a trajectory-level verifier first checks spatiotemporal feasibility and filters unsatisfied travel itinerary, and then the turn-level verifier further validate itinerary's detail consistency with tool responses, enabling efficient and precise reward service. Finally, we propose the reply-augmented reinforcement learning method that enables TP agent to periodically replay from a failure experience buffer, emerging notable agentic capacity. We deploy the trained TP agent in the DiDi Enterprise Solutions application. A three-month online test shows that it achieves 82% accuracy in travel itinerary generation. Comprehensive offline evaluations further demonstrate that DeepTravel enables small-sized LLMs (e.g., Qwen3-32B) to significantly outperform frontier LLMs (e.g., OpenAI o1/o3 and DeepSeek-R1) and existing TP agent frameworks.

cs.AI↗
arXiv2026

Empirical Calibration and Conditional-Reliability Diagnostics for Bearing RUL Prediction under Operating-Regime Shift

Remaining useful life (RUL) estimates support reliability and maintenance decisions only if both point accuracy and prediction intervals remain trustworthy when operating conditions change. Convenient mixed splits can hide that failure. This paper studies the question on a documented 10-bearing PHME subset with time-varying load and speed. Derived load-speed regimes define the held-out evaluation units, while models receive only measured load and speed as context. A calibrated predictive-representation model fuses raw vibration windows, engineered descriptors, and operating context, then forms intervals by empirical residual calibration. Under strict train/validation/calibration/test separation, the model reaches normalized MAE 0.1477, empirical 90% coverage 0.900, and retrospective absolute-step MAE 285.26; a 400-tree random forest reaches 0.1538, 0.871, and 294.57. The results do not show uniform dominance: conditional diagnostics expose non-uniform reliability, including 0.666 coverage in a low-load/high-speed cell, and a post-hoc pooled regime-conditioned residual diagnostic raises that cell to 0.941 only as motivation for future pre-specified conditional calibration. Stress tests further identify raw-channel loss as the largest tested reliability failure mode. The contribution is therefore a bounded reliability-evaluation protocol for the processed 10-bearing subset, with conditional undercoverage and raw-channel loss reported explicitly as failure modes rather than deployment guarantees.

cs.CE↗
arXiv2026

Non-Hermitian topology driven by an identity term: An exactly solvable paradigm

An identity term in the Hamiltonian is conventionally regarded as spectrally inert-it shifts energies but does not alter eigenstate topology. We show that under non-Hermitian skin pumping, this paradigm fails: a momentum-dependent identity term actively deforms the generalized Brillouin zone, thereby challenging established topological criteria that rely on fixed complex contours. Here, by introducing spin-orbit coupling into a Hatano-Nelson chain, we present an exact analytical solution for the entire non-Hermitian eigensystem under open boundary conditions. Our solution reveals how inter-cell spin-orbit coupling, synergizing with this non-trivial identity term, induces topological edge states and robust zero modes in the complete absence of chiral symmetry. This work establishes an exactly solvable paradigm for non-Hermitian topology beyond symmetry protection, and provides a rigorous benchmark for testing topological invariants in systems with momentum-dependent identity terms.

quant-ph↗
arXiv2026

Bgolearn: a Unified Bayesian Optimization Framework for Accelerating Materials Discovery

Efficient exploration of vast compositional and processing spaces remains a major challenge in accelerated materials discovery. Bayesian optimization (BO) provides a principled approach to identify optimal materials with minimal experimentation, but its adoption has been limited by implementation complexity and a lack of domain-specific tools. Here, we present Bgolearn, a versatile Python framework that brings BO to materials research through intuitive interfaces, robust algorithms, and materials-focused workflows. Bgolearn supports single- and multi-objective optimization, multiple acquisition strategies, diverse surrogate models, and uncertainty quantification, enabling effective navigation of complex design spaces. Benchmark studies show that Bgolearn reduces experimental effort by 40-60\% compared with random search, grid search, and genetic algorithms, while achieving comparable or superior solution quality. Its effectiveness is demonstrated across case studies, including the discovery of maximum-elastic-modulus triply periodic minimal surface structures, ultra-high-hardness high-entropy alloys, and high-strength, high-ductility medium-Mn steels, and is further supported by numerous publications. With a modular architecture that integrates seamlessly into existing materials workflows and a graphical interface (BgoFace) that removes programming barriers, Bgolearn establishes a practical, reliable platform for Bayesian optimization in materials science. The software is openly available at https://github.com/Bin-Cao/Bgolearn.

cond-mat.mtrl-sci↗
arXiv2026

RetailBench: Evaluating Long-Horizon Autonomous Decision-Making and Strategy Stability of LLM Agents in Realistic Retail Environments

Large language model (LLM) agents have made rapid progress on short-horizon, well-scoped tasks, yet their ability to sustain coherent decisions in dynamic long-horizon environments remains uncertain. We introduce RetailBench, a data-grounded simulation benchmark for evaluating tool-using LLM agents in single-store supermarket operation. RetailBench models retail management as a partially observable decision process and is designed to support thousand-day-scale simulations. In this environment, agents must manage pricing, replenishment, supplier selection, shelf assortment, inventory aging, customer feedback, external events, and cash-flow constraints. We evaluate seven contemporary LLMs under representative agent frameworks over a 180-day evaluation horizon and compare them with a privileged oracle policy. Results show substantial variation across models: only a small subset survives the full evaluation horizon, and even the strongest LLM runs remain substantially behind the oracle policy in final net worth and sales outcomes. Behavioral analysis attributes these gaps to incomplete evidence acquisition, surface-level decision making, and the lack of a consistent long-horizon policy. RetailBench provides a controlled testbed for studying reliable autonomy in economically grounded long-horizon decision-making.

cs.AI↗
arXiv2026

Fast coherent control of a charge qubit on solid neon with a spin-qubit-compatible resonator

Electrons floating in vacuum provide a clean platform for quantum information processing owing to their isolation from material defects. In particular, electrons on solid neon have emerged as a promising qubit platform because of their potentially long coherence times. Here, toward spin-qubit realization, we couple a single electron on solid neon to a magnetic-field-compatible superconducting NbTiN nanowire resonator. We realize a charge qubit and demonstrate microwave readout and coherent control, with Rabi frequencies up to 76 MHz, an order of magnitude larger than in previous studies. Under strong driving, we observe a qubit frequency shift from nonlinear interactions with the intense microwave field. Deterministic electron trapping at an intended position remains challenging due to solid neon surface roughness; we characterize the electron's position from its differential coupling to distinct electrodes. Although not trapped at an intended position, our estimates indicate that spin-qubit demonstrations remain feasible.

cond-mat.mes-hall↗
arXiv2026

Beyond Scaling Law: A Data-Efficient Distillation Framework for Reasoning

Large language models (LLMs) demonstrate remarkable reasoning capabilities in tasks such as algorithmic coding and mathematical problem-solving. Recent methods have improved reasoning through expanded corpus and multistage training combining reinforcement learning and supervised fine-tuning. Although some methods suggest that small but targeted dataset can incentivize reasoning via only distillation, a reasoning scaling laws is still taking shape, increasing computational costs. To address this, we propose a data-efficient distillation framework (DED) that optimizes the Pareto frontier of reasoning distillation. Inspired by the on-policy learning and diverse roll-out strategies of reinforcement learning, the key idea of our approach is threefold: (1) We identify that benchmark scores alone do not determine an effective teacher model. Through comprehensive comparisons of leading reasoning LLMs, we develop a method to select an optimal teacher model. (2) While scaling distillation can enhance reasoning, it often degrades out-of-domain performance. A carefully curated, smaller corpus achieves a balanced trade-off between in-domain and out-of-domain capabilities. (3) Diverse reasoning trajectories encourage the student model to develop robust reasoning skills. We validate our method through evaluations on mathematical reasoning (AIME 2024/2025, MATH-500) and code generation (LiveCodeBench), achieving state-of-the-art results with only 0.8k carefully curated examples, bypassing the need for extensive scaling. Our systematic analysis demonstrates that DED outperforms existing methods by considering factors beyond superficial hardness, token length, or teacher model capability. This work offers a practical and efficient pathway to advanced reasoning while preserving general capabilities.

cs.LG↗
arXiv2026

A Path-Survival Analytical Framework for SCL Decoding of Polar Codes

A theoretical analysis of CRC-aided successive cancellation list (CA-SCL) decoding for polar codes remains an open problem, despite its widespread practical adoption. While low-density parity-check (LDPC) codes benefit from mature analytical tools, such as density evolution (DE), for predicting the performance of belief-propagation (BP) decoding, similar techniques are not directly applicable to CA-SCL decoding. This limitation stems from the complex path-pruning mechanism inherent in CA-SCL decoding. In this paper, we propose an analytical framework based on a novel path-survival model that captures the evolution of the correct path's rank during decoding. The proposed framework enables efficient prediction of CA-SCL decoding performance without requiring exhaustive list-specific Monte Carlo simulations. Extensive numerical evaluations demonstrate its effectiveness across a wide range of code lengths, code rates, list sizes, and channel models.

cs.IT↗