Searcharxiv⌕ Search

arXiv subjects

Zijian Wang

Publications and source records attributed to Zijian Wang.

At least 19 recordsLinked to original sources
arXiv2026

Heavy Seed Black Hole Growth in Metal-Enriched Halos through Disk-Induced Stellar Disruptions: A Semi-Analytical Modelling

Recent simulations suggest that heavy seed black holes may form in weakly metal-enriched atomic cooling halos, where the supermassive-star progenitor and small-scale stellar fragments emerge nearly coevally. In this picture, the newly born heavy seed is naturally embedded in a metal-enriched Pop~I/II nuclear star cluster rather than in an isolated pristine environment. We investigate whether disk-induced tidal disruption events (TDEs) from these Pop~I/II stars can provide an efficient and sustained growth channel for heavy seed black holes. We construct a semi-analytical model for stellar orbital damping, disk capture, migration, and tidal disruption around a heavy seed black hole, and incorporate the resulting disk-induced TDE contribution into cosmological merger trees with baryonic and metallicity evolution. Heavy seed host halos are selected from atomic cooling halos with $Z\lesssim10^{-3}Z_\odot$ that satisfy either a rapid gas-inflow criterion or a strong Lyman--Werner radiation criterion. We find that disk-induced Pop~I/II TDEs can dominate the early growth of heavy seeds: the median black hole mass grows from $\sim10^4\,M_\odot$ to $\sim10^5\,M_\odot$ within the first $\sim0.1$ Gyr after seed formation, and reaches several $10^5\,M_\odot$ by $\sim0.2$ Gyr. The cumulative mass supplied by TDEs initially exceeds that from gas accretion and remains comparable over the first $\sim200$ Myr. Including disk-induced TDEs shifts the black hole population toward higher masses, increases the abundance of massive black holes at $z\sim9$--10, and produces larger black hole-to-stellar mass ratios. This channel helps alleviate, but does not fully remove, the tension between heavy-seed models and the most extreme high-redshift black hole candidates, suggesting that additional growth mechanisms may still be required.

astro-ph.GA↗
arXiv2026

Projected Sensitivity to Slow Muonphilic Dark Matter with Accelerator Muon Beams

The nature of dark matter (DM) remains one of the most enduring open questions in modern physics, and muonphilic DM has emerged as a promising scenario that complements traditional DM candidates. Following the recently established cosmic-ray muon scattering approach, we investigate the sensitivity for probing slow muonphilic DM with accelerator muon beams. A Geant4-based simulation framework is developed, incorporating the detector geometry from the PKMu muon tomography system and a dedicated elastic $μ$-DM scattering process. The projected sensitivity is found to be largely insensitive to both the beam energy and the transverse beam size when the beam is fully contained within the detector acceptance. For a benchmark beam intensity of $10^5/\rm{s}$, the simulated pure-muon beam surpasses the existing cosmic-ray limit of $1.61\times10^{-17}$ cm$^2$ at $m_{\rm DM}=1$ GeV within approximately 11 seconds. A realistic muon beam phase-space distribution based on simulations for the High Intensity heavy-ion Accelerator Facility (HIAF) is also implemented, yielding projected limits that improve upon the cosmic-ray results by nearly two orders of magnitude in a one-day exposure. These results demonstrate that a beam-muon scattering experiment offers a robust and promising route toward significantly improved sensitivity to slow muonphilic DM.

hep-ex↗
arXiv2026

Rapid Growth of Intermediate-Mass Black Holes through Disk-induced Stellar Disruptions

Dense nuclear star clusters provide unique environments for studying the dynamical interactions between stars and massive black holes. When an accretion disk is present, dissipative star--disk interactions can capture surrounding stars, drive their inward migration, and ultimately lead to disk-induced tidal disruption events\,(dTDEs). The long-term feeding rate from this process, however, cannot be inferred from single-orbit migration estimates alone, as it depends on the coupled evolution of disk capture, collisional relaxation, stellar depletion and replenishment, and physical mergers within the star cluster. In this work, we use high-performance direct $N$-body simulations combined with analytic prescriptions for star--disk interactions to follow this coupled evolution for intermediate-mass black holes\,(IMBHs) with accretion disks embedded in dense stellar clusters. The simulations track the formation of the stellar cusp, the capture of stars by repeated disk crossings, their subsequent orbital damping and migration, and their eventual consumption by the central IMBH. We find that dTDEs can sustain stellar mass supply rates of $\sim10^{-3}\,M_\odot \,\mathrm{yr}^{-1}$, which exceeds the Eddington-limited gas accretion rate for IMBHs with $M_\bullet<10^5\,M_\odot$. These results identify dTDEs as an efficient stellar feeding channel for IMBHs in gas-rich dense stellar systems. As one possible application, this mechanism may help transform $\sim10^3\,M_\odot$ IMBHs into more massive black-hole seeds, provided that compact stellar clusters and accretion disks persist for $>30$ Myr.

astro-ph.GA↗
arXiv2026

Breaking the Code: Security Assessment of AI Code Agents Through Systematic Jailbreaking Attacks

Code-capable large language model (LLM) agents are embedded in software engineering workflows where they can read, write, and execute code, raising "jailbreak" stakes beyond text-only settings. Prior evaluations emphasize refusal or harmful-text detection, leaving open whether agents compile and run malicious programs. We present JAWS-Bench (Jailbreaks Across WorkSpaces), a benchmark spanning three escalating workspace regimes mirroring attacker capability: empty (JAWS-0), single-file (JAWS-1), and multi-file (JAWS-M). We pair it with a hierarchical, executable-aware Judge Framework that tests (i) compliance, (ii) attack success, (iii) syntactic correctness, and (iv) runtime executability to measure deployable harm. Across seven LLM backends from five families, prompt-only attacks in JAWS-0 achieve 61% compliance; 58% are harmful, 52% parse, and 27% run end-to-end. In JAWS-1, compliance reaches ~100% for stronger models with a mean ASR (Attack Success Rate) ~71%; JAWS-M raises mean ASR to ~75%, with 32% runnable attack code. Wrapping an LLM in an agent increases ASR by 1.6$\times$, by overturning initial refusals during planning and tool use. Additional evaluations with SWE-Agent and OpenAI Codex exhibit similar trends, indicating that JAWS-Bench can be reused across multiple agent frameworks. Category analyses identify which attack classes are most vulnerable and deployable, motivating execution-aware defenses and refusal-preserving agent designs.

cs.CR↗
arXiv2026

Pro-Router: Token-Aware Progressive Model Routing with Adaptive Edge-Cloud Collaboration for Efficient Multimodal LLM Inference

The remarkable performance of multimodal large language models (MLLMs) comes at the cost of substantial computational overhead, posing significant challenges to real-time deployment and cost effectiveness. Existing model routing approaches either decide from coarse request-level features alone or spend one or several extra language model passes to inspect the generated response, leaving the token-level uncertainty signals that emerge during generation unused. To address these limitations, we propose Pro-Router, a token-aware progressive model routing method with adaptive edge-cloud collaboration for efficient multimodal LLM inference. Pro-Router employs a two-stage progressive decision mechanism. First, a lightweight prompt pre-scorer module performs rapid pre-screening before token generation begins, guiding apparently simple requests to small models. Second, a token-aware verifier reads the sampling probability distribution of each token the small model generates, estimating the model's confidence in its own output to determine, per request, whether the answer ships or escalates to the cloud-based high-precision model. Furthermore, we design an adaptive edge-cloud serving pipeline that sizes every dispatch to each device's measured service rate, so both the edge and the cloud tiers stay fully utilized without manual parameter tuning and are not impacted by the network latency. Extensive experiments on multiple multimodal benchmark datasets and models demonstrate the effectiveness of Pro-Router. Compared to other methods, it achieves the highest routing accuracy and improves routing speed by more than 10x. Its serving pipeline also reaches more than 75% higher end-to-end throughput than the existing model routing pipeline. Our code is available at https://github.com/xinyuangui2/pro-router.

cs.AI↗
arXiv2026

When Does Visual Generation Help Visual Understanding in Unified Multimodal Models?

Unified multimodal models (UMMs) can perform both understanding and generation, raising a central question: can visual generation improve understanding? Existing evaluations provide mixed evidence, but confound task difficulty, reasoning paradigms, and the closed-loop interaction between generation and understanding. We introduce VGAU-Diag, a fine-grained evaluation framework for vision generation-assisted understanding. It stratifies samples by difficulty, enables unified evaluation of multiple reasoning paradigms, and uses Oracle-Assisted Reference Protocols. Our analysis shows that generated visual aids help on easier instances but become unreliable as reasoning complexity increases. Oracle-assisted diagnosis further reveals that the main bottleneck often lies on the visual-understanding side rather than the visual-generation side, as current UMMs struggle to leverage even faithful visual aids. We also show that effective visual generation should target visual-understanding bottlenecks rather than add more reasoning steps, and identify a three-stage transition from task-irrelevant noise, to misleading plausible guidance, and finally to useful assistance. These findings would be useful to guide the development of better UMMs.

cs.CV↗
arXiv2026

Mitigating Database Leakage in RAG Systems with Keyword-Grounded Fact Substitution

Retrieval-Augmented Generation (RAG) has emerged as a powerful paradigm for combining large language models (LLMs) with external knowledge sources. However, RAG systems remain vulnerable to prompt injection attacks, which may mislead the retriever or generator to expose sensitive database contents. To address this issue, we propose KFS-RAG, a defense that mitigates information leakage by reformulating the retrieved context. Specifically, our method first identifies a small set of influential keywords from the retrieved context via an attention rollout plus a causal perturbation mechanism. These keywords are then used to guide an auxiliary LLM to generate a compact set of keyword-grounded facts from the retrieved passages. Finally, the original context is substituted with these curated facts, ensuring that the generator operates on sanitized evidence rather than the raw retrieved text. Experimental evaluations demonstrate that KFS-RAG significantly reduces the risk of database leakage under injection attacks while maintaining response accuracy and relevance. This work highlights a practical pathway toward building secure and trustworthy RAG systems.

cs.CL↗
arXiv2026

GS-Voxel: Fitting-Free Structured Latents for Large-Scale 3DGS Generation

Many scalable latent 3D generators operate on structured tensors, whereas pre-optimized 3D Gaussian Splatting (3DGS) reconstructions are unordered, spatially irregular, and vary widely in primitive count. We present GS-Voxel, a fitting-free structured latent framework, and evaluate it for large-scale aerial 3D Gaussian scene generation. GS-Voxel deterministically converts a compatible pre-optimized 3DGS reconstruction into sparse active voxels without additional per-scene optimization, retaining the sub-voxel positions and rendering attributes of the selected primitives. A GS-specific factorized VAE then separately encodes voxel geometry and local Gaussian attributes into sparse 3D latents whose size grows with the number of occupied voxels rather than being limited by a fixed scene-wide primitive count. We train image-conditioned flow models in the GS-Voxel latent space to generate aerial 3DGS scenes. A key application enabled by GS-Voxel is large-area scene generation: overlap-aware tiled inference extends synthesis beyond a single training crop conditioned on satellite-view images. Our results show that GS-Voxel provides structured latents for pre-optimized aerial 3DGS reconstructions, with latent capacity that grows with the number of occupied voxels.

cs.CV↗
arXiv2026

StellaVLA: In-Context Structured Demonstration for Generalizable Vision-Language-Action Models

Vision-Language-Action (VLA) models can follow instructions and manipulate objects, but their performance often collapses out of distribution (OOD), when the scene, viewpoint, or object differs from training. Adapting to each new situation typically requires collecting more data and fine-tuning. We present StellaVLA, a framework that instead adapts at test time by conditioning on a single retrieved demonstration. The key idea is to move beyond imitating what an expert did and instead convey why: an automated offline pipeline converts each raw trajectory into a structured demonstration, e.g., a task plan, sub-goal descriptions, and verbalized 3D motion, at zero human-annotation cost. Provided as in-context guidance, this structured demonstration lets the policy reason about the task rather than mimic a pixel trajectory, which also makes it transferable across embodiments (real-robot, human-hand, or XR demonstrations). A parallel dual-training design internalizes this reasoning during training through a joint action-and-language objective, while inference uses the action expert alone, preserving real-time, high-frequency control with no added latency. On the VLA-Arena leaderboard(Aug 1, 2026), StellaVLA ranks first with an overall score of 0.63, versus 0.44 and 0.22 for the strong prior models ($π_{0.5}$ and LingBot-VLA), and it further leads on LIBERO with 98.8% average success rate and LIBERO-Plus with 85.1% success rate. Our real-robot benchmark demonstrates that StellaVLA can use both human/robot demos and human-to-robot (XR) demos as in-context structured demonstration to help VLA model adapt to OOD tasks.

cs.RO↗
arXiv2026

PURPOSE: Poisoning Conflict Resolution in RAG via Proxy-Fact-Grounded Updates

In Retrieval-Augmented Generation (RAG), post-retrieval conflict resolution arbitrates among noisy or contradictory retrieved passages. However, the robustness of this safeguard against knowledge poisoning has not been adequately studied. Existing black-box poisoning methods all assert the target answer in frontal contradiction with what the resolver treats as settled, the very signal these methods are built to detect. We propose PURPOSE, a strict black-box poisoning attack that reframes the injection as an update that minimizes conflict, rather than as a counter-claim. PURPOSE extracts query-related facts approximating the resolver's possible reference, then grounds a pivot event in them to keep the injection consistent with what the resolver might verify while steering the generator toward the target answer. Across three QA benchmarks, five generators, and three conflict-resolution methods, PURPOSE attains the highest attack success rate (ASR) in 35 of 45 settings and exceeds the strongest prior attack with +9.7 mean ASR points. These results show that our poisoning method is effective against conflict resolution in RAG and identify non-contradicting injection as a practical mode to enhance poisoning attack.

cs.CR↗
arXiv2026

LAVE: Latent Visual Evidence-Enhanced Planning for Video Tool-use Agents

Long-video understanding requires models to efficiently acquire and reuse sparse visual evidence from long and redundant video streams. Recent video tool-use agents address this challenge by iteratively invoking visual Tools at different temporal scales, but their Tool-Planner communication typically relies on textual observations. Such text-only interfaces provide lossy summaries of Tool computations, causing previously computed visual evidence not verbalized to be discarded and unavailable for subsequent planning. We identify this limitation as the Tool observation bottleneck and propose Latent Visual Evidence-Enhanced Planning (LAVE), a training-free framework for reusing latent visual evidence from completed Tool calls. LAVE introduces a dual-channel observation interface: the visible channel preserves the original textual trajectory, while the latent channel stores pre-verbal visual updates with their Tool roles, source-frame timestamps, and visual locations. During planning, LAVE retrieves evidence relevant to the current Planner state but not covered by textual observations, and integrates it through bounded timestamp-aligned latent updates with entropy-constrained frame-time routing. This enables video agents to reuse existing visual computation without additional training, frame replay, or modifications to the original orchestration. Extensive experiments on Video-MME, LongVideoBench, and CG-Bench show that LAVE consistently improves video tool-use agents across backbones. Under a comparable frame budget, LAVE improves the Video-MME overall score by 3.76 points over the strongest baseline, demonstrating the effectiveness of latent visual evidence reuse for multi-step video-agent planning.

cs.CV↗
arXiv2026

Externalizing Research Synthesis and Validation in AI Scientists through a Research Harness

AI systems can increasingly automate scientific workflows, but the reasoning that links prior evidence, generated ideas, experiments and final claims often remains implicit inside model inference. Here we introduce Xcientist, a research harness that externalizes research synthesis and experimental validation into inspectable, contract-governed processes. Xcientist organizes literature evidence, idea states, implementation plans, ablation records and repair traces as persistent research artifacts, so that generated mechanisms can be grounded, executed, tested and revised without losing their evidential basis. We identify claim drift as a failure mode of automated research, where runnable artifacts no longer support the mechanism originally claimed. Across training-free memory systems, graph-structured traffic forecasting and multi-scale physics-informed neural networks, Xcientist preserves traceable trajectories from problem formulation to mechanism design, validation and bounded revision. These results suggest that AI scientists should be evaluated not only by their final artifacts, but by whether their synthesis and validation processes remain attributable, inspectable and scientifically accountable.

cs.AI↗
arXiv2026

Production of lepton-flavor-violating scalars through resonant positive-muon annihilation on atomic electrons

We investigate an invisible lepton-flavor-violating scalar $ϕ$ with exclusive $e-μ$ couplings and study its resonant production via $μ^+e^-\toϕ$ in fixed-target experiments. Since the effective center-of-mass energy is determined by the momentum of the initial-state bound electrons, atomic effects can significantly affect the resonance behavior. We therefore employ relativistic bound-state electron wave functions to calculate the production cross section and reveal a material-dependent broadening of the resonance lineshape. For the proposed HIAF experiment, fewer than one day of data taking ($6\times10^{10}$ MOT) can probe couplings at the $10^{-5}$ level at 90\% confidence level near resonance, demonstrating that high-intensity muon fixed-target experiments provide a powerful complementary probe of lepton-flavor violation.

hep-ph↗
arXiv2026

PolyInterview: An LLM-based Platform for Immersive Mock Interview Practice with Comprehensive Multimodal Assessment

Preparing for job interviews is important for securing desired positions, yet realistic practice remains difficult to access: real interviews are infrequent, expert mock coaching is costly, and self-practice offers neither adaptive dialogue nor structured assessment. Existing systems typically address only parts of this need through fixed question sequences, limited communication channels, or feedback with little supporting evidence. We present PolyInterview, an LLM-based platform for immersive mock interview practice with comprehensive multimodal assessment. PolyInterview uses the target job description and CV to generate questions tailored to the role and candidate, conducts multi-turn spoken interviews with a lip-synced digital human interviewer that asks answer-aware follow-up questions, and evaluates response content, vocal delivery, and non-verbal behavior. Four parallel evaluators produce 13 behavior-level features that are aggregated into 10 assessment aspects and two competency tracks. Guided by the KSA and STAR frameworks, the report links each score to behavioral evidence and actionable recommendations. PolyInterview is publicly accessible. Its current all-account snapshot contains 101 accounts, 1,564 interview sessions, 7,665 generated questions, and 1,422 five-stage question sets. Generated questions are more closely aligned with their matched job description than with cross-role job descriptions in 93.7% of sessions. An evaluation by ten experts found strong question plans and actionable feedback.

cs.CL↗
arXiv2026

Physical aging of glasses of an organic semiconductor

All glasses, including organic semiconductor glasses, are non-equilibrium materials whose properties will change with time. This physical aging process is poorly understood for organic semiconductors, hindering the rational design of highly durable devices. In this study, we investigated the volume and enthalpy recovery processes in both thin films and bulk glasses of N,N'-Bis(3-methylphenyl)-N,N'-diphenylbenzidine (TPD). Our results revealed that volume recovery kinetics exhibit negligible dependence on film thickness for liquid-cooled TPD films between 400 nm and 100 nm. Additionally, the volume recovery process in TPD films was strongly coupled to the enthalpy recovery observed in bulk TPD glasses during annealing near the glass transition temperature. Remarkably, TPD films prepared by physical vapor deposition at room temperature demonstrated exceptional resistance to physical aging, with an aging rate approximately one order of magnitude lower than that of their liquid-cooled counterparts. These results not only enhance our understanding of the non-equilibrium dynamics in amorphous systems but also offer valuable insights for the design of next-generation organic devices with significantly improved stability and durability.

cond-mat.mtrl-sci↗
arXiv2026

FedSPM: Routing-Enabled Federated Learning under Dual Heterogeneity via Semiparametric Mixture

Routing-prediction federated learning has emerged as a new paradigm that reframes inter-client heterogeneity as a resource for system-level intelligence: at inference time, the server routes each external query to the best-matched client for prediction. Existing approaches, however, typically treat each client as internally homogeneous, overlooking latent subpopulations within local data. For example, patients with the same diagnosis at one hospital may exhibit morphologically distinct disease subtypes. The coexistence of inter-client and intra-client heterogeneity, which we call dual heterogeneity, can impair both routing and prediction. To address this challenge, we propose FedSPM, a routing-enabled semiparametric mixture framework that represents each client using client-specific latent components. Each component combines a predictive distribution for classification with a feature distribution for routing. To flexibly model feature distributions while effectively sharing information across clients, FedSPM models their density ratios relative to a common nonparametric measure estimated via empirical likelihood. We develop a federated expectation-maximization algorithm that optimizes a tractable surrogate and prove convergence of the exact profiled objective at the standard $\mathcal{O}(1/\sqrt{T})$ rate when the surrogate errors are properly controlled. Experiments on controlled benchmarks and real-world medical data demonstrate consistent improvements in routing and prediction under dual heterogeneity. Code is available at https://github.com/zijianwang0510/FedSPM.

cs.LG↗
arXiv2026

Humans are Missing from AI Coding Agent Research

Recent progress in AI coding agent research has led to rapid improvements in agents' ability to autonomously perform complex software engineering tasks, from editing large codebases to executing long-horizon development workflows. As these systems make strides, however, the primary bottleneck to practical usefulness increasingly shifts away from pure task-solving capability, and toward challenges in how users communicate with, supervise, and trust agents. In this position paper, we argue for a reorientation from autonomous to human-centered coding agents: systems designed not only to complete tasks, but to collaborate effectively with people. We identify four core interaction-level dimensions that characterize the human-agent task-solving loop: task alignment, verifiability, steerability, and adaptability. Finally, we outline concrete research directions to advance these dimensions, including user-involved coding environments, comprehensive verification mechanisms, and principled measures of human-agent interaction quality.

cs.HC↗
arXiv2026

PosterHarness: Turning Scientific Poster Generation into an Auditable Instruction-Following Benchmark

Text-rich image models can now design poster-scale layouts, but we lack ways to measure whether they honor scientific communication contracts: legible labels, prescribed aspect ratios, and -- above all -- abstaining from fabricated scientific figures. We present POSTERHARNESS, an auditable harness reframing poster generation as measurable instruction-following tasks, with a pilot benchmark and failure taxonomy. POSTERHARNESS uses a placeholder-first contract to separate two jobs models otherwise conflate. The model performs visual-summary design: typography, reading path, color, and background -- but never draws data-bearing figures. Every figure region must be an empty labeled placeholder; a deterministic compositor inserts real source-paper figures at detected coordinates. This makes properties measurable: placeholder count and ID accuracy, blankness, aspect-ratio compliance, abstention from synthesized graphics, public-text hygiene, and source-figure provenance -- with failures logged as explicit rejections, not hidden in plausible-looking output. We instantiate the harness on 12 papers (6 HEP, 6 AI/ML-adjacent) and report three findings. (i) A counterfactual probe shows the placeholder contract drives VLM-counted synthesized figures from 34 to 0 across three papers. (ii) A failure taxonomy identifies blocking contracts: placeholder geometry, placeholder QA, template critic, and public text. (iii) Comparison with Paper2Poster shows a trade-off: PosterHarness yields higher-resolution artifacts, lower white-canvas fraction, and stronger VLM visual preference; the deterministic baseline retains slightly more PosterQuiz-style information and runs faster. We report this as regime characterization, not a superiority claim. All artifacts, prompts, manifests, and audit scripts are released as a reusable evaluation component.

cs.CV↗