arXiv subjects

Jiaqi Wang

Publications and source records attributed to Jiaqi Wang.

At least 55 records · Page 3Linked to original sources

arXiv2026

ConvergeWriter: Data-Driven Bottom-Up Article Construction

Large Language Models (LLMs) have shown remarkable prowess in text generation, yet producing long-form, factual documents grounded in extensive external knowledge bases remains a significant challenge. Existing "top-down" methods, which first generate a hypothesis or outline and then retrieve evidence, often suffer from a disconnect between the model's plan and the available knowledge, leading to content fragmentation and factual inaccuracies. To address these limitations, we propose a novel "bottom-up," data-driven framework that inverts the conventional generation pipeline. Our approach is predicated on a "Retrieval-First for Knowledge, Clustering for Structure" strategy, which first establishes the "knowledge boundaries" of the source corpus before any generative planning occurs. Specifically, we perform exhaustive iterative retrieval from the knowledge base and then employ an unsupervised clustering algorithm to organize the retrieved documents into distinct "knowledge clusters." These clusters form an objective, data-driven foundation that directly guides the subsequent generation of a hierarchical outline and the final document content. This bottom-up process ensures that the generated text is strictly constrained by and fully traceable to the source material, proactively adapting to the finite scope of the knowledge base and fundamentally mitigating the risk of hallucination. Experimental results on both 14B and 32B parameter models demonstrate that our method achieves performance comparable to or exceeding state-of-the-art baselines, and is expected to demonstrate unique advantages in knowledge-constrained scenarios that demand high fidelity and structural coherence. Our work presents an effective paradigm for generating reliable, structured, long-form documents, paving the way for more robust LLM applications in high-stakes, knowledge-intensive domains.

cs.CL↗

arXiv2026

Robust Bimanual Vision-Language-Action Models via Embarrassingly Simple Modality Masking

Query-based Vision-Language-Action (VLA) models offer low-latency inference that is attractive for bimanual robotic manipulation, but we observe that they can still exhibit discontinuous actions and execution failures in complex dual-arm tasks. We hypothesize that unstable multi-view and language fusion is one contributing factor in these failures, often coinciding with attention spreading to distracting regions. To improve robustness, we introduce the Modality Masking Mechanism (M3), an embarrassingly simple, training-only strategy that requires no architectural changes or large-scale robot pretraining. M3 stochastically masks subsets of modality channels during training, exposing the policy to controlled partial observations and encouraging it to rely less on distracting cues and more on evidence that remains reliable. We evaluate M3 on ten bimanual tasks from RoboTwin 2.0 and on three long-horizon real-world tasks. Compared with the Adapter baseline, M3 improves average success by 21.7% in the Clean setting and 11.4% in Clean2Rand, where policies are trained on clean demonstrations and evaluated on randomized scenes, while also improving averaged real-world full-task success by over 30%. These results suggest that structured training-time masking is a practical way to improve the robustness of query-based VLA policies for bimanual manipulation.

cs.RO↗

arXiv2026

OrthoSkillVLA: Continual Skill Learning via Gradient-Informed Skill Subspace Adaptation

Pretrained Vision-Language-Action models provide a strong foundation for robot learning, but sequentially adapting them to diverse skills can perturb the representations and velocity mappings used by previous skills, leading to catastrophic forgetting. Architecture-based approaches improve retention by isolating skills but lead to increased inference footprint. Recent subspace-constrained methods restrict parameter updates in an orthogonal subspace to minimize interference but impose a unified constraint on the entire model. We analyze the distinct roles of internal VLA components and identify two VLA-specific challenges. First, the VLM maintains broad semantic representations, making it vulnerable to capacity exhaustion, whereas the ActionHead refines semantics into localized velocity patterns that are highly sensitive to perturbations. Second, the final velocity decoder serves as a readout layer. Freezing it forms an output-stage expressivity bottleneck, while updating it risks overwriting previous velocity mappings. To this end, we propose OrthoSkillVLA, a parameter-efficient framework for continual skill learning in pretrained VLA models without demonstration replay. Given the representation heterogeneity, we impose separate subspace constraints on the VLM and ActionHead, preserving reusable semantic capacity while protecting localized velocity patterns. For the output layer, we introduce a lightweight feature-aware MoE decoder, where each skill is allocated a compact expert and a training-free router selects the expert according to feature-space affinity. Extensive simulated and real-world evaluations, together with ablations, demonstrate that OrthoSkillVLA better preserves prior skills while acquiring new ones.

cs.RO↗

arXiv2026

Margin-Regularized Structured Semantic Alignment for Brain-Language Correspondence

With the rapid advancement of large language models, brain-language decoding has achieved remarkable progress. However, it remains unclear whether decoded content genuinely reflects neural representations or is largely reconstructed by the language model itself. This ambiguity limits interpretability and hinders the investigation of intrinsic brain-language correspondence. To address this challenge, we propose MD-SigLIP. This margin-regularized structured semantic alignment framework directly aligns brain embeddings with text embeddings in a shared semantic space, enabling retrieval-based decoding. This formulation enables explicit modeling of the correspondence between neural representations and language semantics. Building upon duplicate-aware sigmoid contrastive learning, we introduce a listwise margin-regularized term that enforces structured ranking constraints between positive semantic clusters and negative samples. By modeling multi-positive semantic structure and margin-based ordering simultaneously, the method captures the manifold organization of language embeddings reflected in neural signals. Experiments demonstrate state-of-the-art retrieval performance under both full-vocabulary and subset evaluation settings.

cs.CL↗

arXiv2026

HPSD: Hybrid-Policy Self-Distillation for Text-Image-to-Video Diffusion Models

Text-Image-to-Video (TI2V) models are an emerging unified architecture, where a single model simultaneously supports text-to-video (T2V) and image-to-video (I2V) generation. Given a high-quality first frame or a detailed textual prompt, TI2V models unlock substantially better visual quality than their T2V mode, raising a natural question: can the capability elicited by such privileged conditions be internalized into the model's own base generation ability? A common approach toward this goal is model self-distillation. However, the most straightforward solution, supervised fine-tuning, follows an off-policy strategy: its supervision is confined to teacher-generated endpoints from a fixed offline distribution rather than student-visited states, lacking precise correction tailored to the evolving policy. Recent on-policy distillation methods instead suffer from condition-state mismatch, where supervision is steered toward the given first frame instead of the student's actual content, misleading the correction. To achieve self-distillation that absorbs the teacher's privileged prior while retaining precise policy correction, in this work, we propose Hybrid-Policy Self-Distillation (HPSD), a novel self-distillation framework where a single TI2V model acts as both teacher and student under different conditions: the teacher operates in TI2V mode with a high-quality first frame and an enhanced prompt, while the student runs in the base T2V mode with only the vanilla prompt. Specifically, the student inherits off-policy teacher trajectory points as anchors, locally refines them toward its own policy, and finally receives velocity-level supervision on these self-generated roll-outs. Extensive experiments demonstrate that HPSD significantly improves T2V performance while also delivering notable TI2V gains, effectively strengthening the model's base generation ability.

cs.CV↗

arXiv2026

A novel projected shell model method for nuclear level density

We introduce a shell model method for calculating nuclear level density (NLD) generally applicable for arbitrarily heavy deformed nuclei. The method is novel because of our use of physical guidance for the construction of its configuration space and the computational breakthrough with the Pfaffian algorithm. Taking a well deformed even-even $^{164}$Dy nucleus as an example, we solve exactly the eigenvalue equation to obtain a large ensemble of eigenstates of angular momentum and parity. Our results indicate a potential need to revise some common understanding of the structural behavior in the pair-breaking region where the structure changes in NLD of the compound nuclei would be sensitive for cross-sections of neutron capture.

nucl-th↗

arXiv2026

Projected shell model description of nuclear level density: Collective, pair-breaking, and multiquasiparticle regimes in even-even nuclei

There is overwhelmingly experimental evidence indicating that excited nuclear states are dominated by quasiparticle (qp) excitations, which form many-body configurations with broken nucleon-pairs from different orbitals. By using these multi-qp states as building blocks for a shell-model basis, we propose a novel shell-model method to calculate the nuclear level density (NLD) in deformed nuclei. The shell-model diagonalization with two-body residual interactions yields a large ensemble of eigenstates of angular momentum and parity. We demonstrate that NLD as a statistical quantity depends sensitively on the structure of deformed single-particle states. As the first example to introduce this method, we take a well-deformed rare-earth nucleus, $^{164}$Dy, for which NLD has been studied extensively by the Oslo method. By a quantitative comparison with discrete levels from spectroscopic measurements, we show that while the pronounced stepwise structure in the low-energy NLD curve can be understood as the collective excitation and nucleon-pair breaking, the exponential growth of levels in the higher-energy NLD can be described by the combination of the broken-pair states, subject to the Pauli principle. According to the nature of NLD with increasing excitation, we divide the entire NLD curve into (1) collective regime, (2) pair-breaking regime, and (3) multi-qp regime. We discuss the formation mechanism and characteristic features of NLD for the three regimes. In addition, the parity dependence and angular-momentum dependence in NLD are investigated with a strong emphasis on the structure effect.

nucl-th↗

arXiv2026

Nuclear level density studied in odd-mass nuclei in the framework of the projected shell model

In a recent article [Phys. Rev. C 108, 034309 (2023)], we proposed a projected shell model method for the calculation of nuclear level density (NLD) in deformed even-even nuclei. The current article presents the subsequent study of NLDs in odd-mass nuclei as well as a comparative analysis between our calculated NLDs in adjacent even-even and odd-A systems. Since one nucleon in the odd-mass system remains blocked from participating in the pair formation, resulting in a weakened pairing (assessed by a smaller BCS pairing gap), pronounced differences between the NLDs in an odd-mass (both even-odd and odd-even) nucleus and its immediate even-even neighbour have been found. In general, the structure-dominated variations, which were found to be prominent in the even-even NLD at low energies, are greatly suppressed in the odd-mass systems. Specifically, from excitation energy as low as 2 MeV, the calculated densities of odd-parity and even-parity levels in odd-mass nuclei show an equal division signaling faster attainment of the statistical behavior. Nuclear level-spin distributions of both parities have been seen to adopt a regular Gaussian shape earlier than that found in the even-even system. Moreover, the pleasant property of our shell-model results, that each of our calculated levels is an eigenstate of angular momentum, allows us to extract the values of the energy-dependent dispersion $σ$ of Ericson's spin-distribution formula and plot $ρ(E, I, π)$, the energy-, spin-, and parity-dependent level density.

nucl-th↗

arXiv2026

FailForge: Distilling Procedural Competence from Persistent Failures into Code Agents

Rejection sampling fine-tuning (RFT) is widely used to train code agents by generating trajectories on verifiable software engineering tasks, retaining those that pass the tests, and fine-tuning on the successful rollouts. However, even strong code agents repeatedly fail on a substantial fraction of such tasks, and standard RFT simply discards these failures. The discarded samples are precisely the hardest and most informative ones, drawn from verifiable instances that are costly to curate. Stronger base models may reduce the number of failures, but the remaining hard cases still define the frontier for further improvement. We propose FailForge, an agentic framework that converts failed rollouts into training signal. For each failed instance, an agent diagnoses the failure from error feedback and execution traces, distills the diagnosis into a concise and actionable skill, and injects the skill into the agent context for a guided second attempt. Trajectories that succeed under skill guidance are folded back into the RFT corpus. Crucially, the skill is removed at training time, so the model internalizes the recovered behavior rather than relying on external hints at inference. FailForge recovers over 26% of previously failed instances at marginal additional cost, and training Qwen3.5-4B on the augmented corpus improves the SWE-bench Verified resolve rate by 6.6 points over a strong RFT baseline, with gains concentrated on the hardest problems.

cs.AI↗

arXiv2026

UniREditBench: A Unified Reasoning-based Image Editing Benchmark

Recent advances in multi-modal generative models have driven substantial improvements in image editing. However, current generative models still struggle with handling diverse and complex image editing tasks that require implicit reasoning, underscoring the need for a comprehensive benchmark to systematically assess their performance across various reasoning scenarios. Existing benchmarks primarily focus on single-object attribute transformation in realistic scenarios, which, while effective, encounter two key challenges: (1) they largely overlook multi-object interactions as well as game-world scenarios that involve human-defined rules, which are common in real-life applications; (2) they only rely on textual references to evaluate the generated images, potentially leading to systematic misjudgments, especially in complex reasoning scenarios. To this end, this work proposes UniREditBench, a unified benchmark for reasoning-based image editing evaluation. It comprises 2,700 meticulously curated samples, covering both real- and game-world scenarios across 8 primary dimensions and 18 sub-dimensions. To improve evaluation reliability, we introduce multimodal dual-reference evaluation, providing both textual and ground-truth image references for each sample assessment. Furthermore, we design an automated multi-scenario data synthesis pipeline and construct UniREdit-Data-100K, a large-scale synthetic dataset with high-quality chain-of-thought (CoT) reasoning annotations. We fine-tune Bagel on this dataset and develop UniREdit-Bagel, demonstrating substantial improvements in both in-domain and out-of-distribution settings. Through thorough benchmarking of both open-source and closed-source image editing models, we reveal their strengths and weaknesses across various aspects.

cs.CV↗

arXiv2026

EpiBench: Can LLMs Understand Epitopes for Antibody Drug Discovery?

Epitopes determine where antibodies bind antigens and shape downstream therapeutic properties such as functional blockade and escape resistance, making epitope understanding central to antibody drug discovery. Although large language models (LLMs) have shown strong biomedical reasoning ability, it remains unclear whether they can infer epitope information directly from antigen and antibody sequences. Existing epitope resources typically focus on isolated prediction tasks or rely on specialized structural settings, while general protein benchmarks do not evaluate epitope-centered decisions across the antibody development workflow. To address this gap, we introduce EpiBench, a closed-book, sequence-based, and automatically scorable benchmark for evaluating epitope reasoning in LLMs. EpiBench contains 1,609 curated samples grounded in structural antibody--antigen contacts, curated functional B-cell assays, and deep mutational scanning escape measurements. It covers five connected tasks: targetable region discovery, antibody-conditioned epitope identification, epitope binning, functional epitope assessment, and antibody escape assessment, with controlled sampling to reduce shortcut-based evaluation artifacts. We evaluate nine general-purpose LLMs and analyze their behavior through task-specific baselines, antigen length stratification, explicit-reasoning comparison, and failure-mode inspection. The results show that current LLMs capture partial epitope-related signals but remain limited in antibody-specific sequence grounding, long-context residue localization, and biologically grounded reasoning. Therefore, EpiBench provides a diagnostic testbed for measuring and improving sequence-aware biomedical LLMs toward reliable LLM-assisted antibody discovery.

cs.CL↗

arXiv2026

SearchAuditor: Auditing and Attributing Failures in Long-Horizon Search Agents

Deep search agents tackle challenging questions through long-horizon web interactions, a process that is both complex and fragile: small reasoning errors may propagate through long, noisy trajectories into fluent but incorrect answers. Diagnosing such failures is difficult, requiring the manual inspection of extremely long execution traces, which could be beyond human capacity. We therefore introduce SearchAuditBench, a benchmark that evaluates whether LLM auditors can localize, attribute, and repair these failures, thereby reducing the human burden. SearchAuditBench comprises 1,243 failed trajectories, averaging 73.1 messages and 65.1K tokens, collected from eight open-weight models on five deep-search benchmarks, each expert-annotated with the critical error step, a search-specific root cause, and a reference repair with grading rubrics. We further propose SearchAuditor, a multi-perspective auditing framework that effectively localizes, attributes, and repairs search-agent failures through evidence-grounded adjudication. Experimental results show that even the strongest baseline, when powered by a frontier model like GPT-5.5, attains only a 26.6% end-to-end pass rate. In contrast, our SearchAuditor consistently outperforms all baselines across different frontier models, achieving an end-to-end pass rate of 32.3%, and resuming failed runs with its repairs enables agents to better recover from errors.

cs.AI↗

arXiv2026

Closing the Gap: Efficient Algorithms for Discrete Wasserstein Barycenters

Given a collection of probability measures, the Wasserstein barycenter problem seeks a probability measure that minimizes the weighted sum of their squared type-$2$ Wasserstein distances. We study the discrete setting, in which every input measure has finite support; this setting arises frequently in machine learning and operations research. Since computing a discrete Wasserstein barycenter is NP-hard, we develop approximation algorithms with provable performance guarantees. The best previously known polynomial-time guarantee is a factor of $2$. We improve this bound by proposing a polynomial-time approximation scheme (PTAS) that generalizes the existing $2$-approximation: for any prescribed $α\in(0,1]$, the proposed algorithms return a $(1+α)$-approximate barycenter in time polynomial in $(nk)^{1/α}$ and $d$. We provide both randomized and deterministic constructions and derive a sharper guarantee when the input measures are equally weighted. Numerical experiments on synthetic and real datasets show that the algorithms are computationally practical and produce near-optimal barycenters.

math.OC↗

arXiv2026

JoyAI-Talker: Full-Duplex Speech Interactive Large Model Built for Empathetic Voice Agents

We present JoyAI-Talker, a full-duplex speech dialogue system that delivers robust foundation model capabilities while empowering empathetic interaction and voice agent intelligence. JoyAI-Talker adopts a modular Thinker-Talker architecture and further implements a unified speech-text joint training pipeline to mitigate the common "cognitive degradation" bottleneck, thereby largely preserving the model's core textual reasoning, STEM, and logical capabilities while extending them to speech-based interaction. For expressive speech synthesis, the Talker module employs a text-controllable generation paradigm that enables natural-language instructions to flexibly control vocal attributes and localized paralinguistic events, such as laughter and sighs, supporting more expressive and fine-grained speech responses. To enhance conversational empathy, we introduce the Persona-Adaptive Empathetic Response (PAER) framework. PAER employs a hierarchical cognitive pipeline to extract non-verbal speaker cues, such as gender, age, and emotional state, from raw input audio, incorporate them into the Thinker's CoT reasoning, and generate context-adaptive responses that align semantically appropriate text with fine-grained control over utterance-level expressiveness and localized paralinguistic events, including sighs, speaking rate, and volume. We further integrate Joy-Duplex, a state-driven, plug-and-play full-duplex framework that functions as an efficient gating engine for real-time turn control. Extensive evaluations show that JoyAI-Talker achieves highly competitive performance on foundational T2T and S2T benchmarks. In full-duplex evaluation, the system reaches a high response rate of 0.88 under user interruptions while maintaining an extremely low false-trigger rate under background speech, demonstrating its readiness for fluid and natural speech dialogue.

cs.SD↗

arXiv2026

Intraday Gas Fee Heterogeneity on Ethereum: Evidence from Operational Firms

Ethereum's EIP-1559 fee mechanism was designed under the assumption of homogeneous, myopic agents responding to a single congestion signal. We examine how this assumption interacts with the heterogeneous demand structure of real-world Ethereum users. Analyzing 62,142 confirmed transactions from seven operational firms across seven industries (January--March 2026), we document significant intraday gas-fee variation: fees peak at hour~12 UTC (7\,AM ET, $\hatβ_{12}=\$0.054$ above the U.S.\ evening baseline, $p<0.001$) and are associated with periods of elevated speculative-arbitrage activity. Operational firms exhibit heterogeneous scheduling responses moderated by transaction deferrability and gas intensity. Residual cost floors, i.e. the gap between observed expenditure and the counterfactual under perfect off-peak scheduling, range from 40.7\% to 92.5\% of actual expenditure, and persist even during the lowest-cost hours ($h\in\{20,21,22,23\}$ UTC, 3--6\,PM ET). We introduce an On-Chain Scheduling Matrix that maps firms to four scheduling regimes as a practical framework for managing gas-fee exposure under the current mechanism.

econ.EM↗

arXiv2026

Dual-Mode SERS and Colorimetric Sensor for Lung Cancer VOC-Biomarker Detection Using Hydrogel Patches

Hexanal, a volatile organic compound (VOC), is a potential biomarker for the early detection of lung cancer. In this study, we developed a dual-mode flexible biosensor that integrates surface-enhanced Raman scattering (SERS) and colorimetric detection for the quantitative analysis of hexanal in human exhaled breath. The biosensor employs Ag nanocubes wrapped with Co-Ni layered double hydroxide (AgNCs@Co-Ni LDH) as a functional matrix, offering both superior SERS enhancement and efficient VOC adsorption properties. To achieve selective detection, AgNCs@Co-Ni LDH were incorporated into agarose hydrogels along with 3-methyl-2-benzothiazolinone hydrazone (MBTH). The resulting MBTH-AgNCs@Co-Ni LDH/hydrogel-patch facilitate the oxidation of hexanal, enabling simultaneous colorimetric and SERS signal generation, while producing acrizine, a blue-colored reaction product. This hydrogel-based dual-mode sensing platform exhibits high selectivity, excellent stability, and precision in SERS-based hexanal detection. The detection limit for the SERS method was determined to be as low as 3.34x10-13 M. Furthermore, developed and optimized compact CNN-based multi-terminal intelligent recognition system for enhanced hydrogel-patch detection through AI-driven, portable, and real-time colorimetric analysis. Therefore, this work not only enables the effective detection of hexanal in the exhaled breath of suspected lung cancer patients, underscoring its potential for early lung cancer screening, but also establishes a foundation for the development of multimodal hydrogel biosensors for broader applications in disease diagnosis.

physics.med-ph↗

arXiv2026

AlphaCrafter: Harnessing Multi-Agent Workflows for Cross-Sectional Quantitative Trading

Quantitative trading agents have demonstrated substantial promise in automating factor discovery, signal aggregation, and portfolio execution. However, existing agent-based trading systems predominantly rely on loosely specified natural-language workflows, leading to opaque reasoning processes, inconsistent behaviors across foundation models, and limited controllability and verifiability, all of which introduce significant risks in financial decision-making. To address these limitations, we propose AlphaCrafter, a multi-agent framework built upon a structured agent harness. Instead of treating agent behavior as unconstrained prompt execution, AlphaCrafter encapsulates each agent within programmable policy specifications that integrate procedural workflows, execution constraints, and explicit verification mechanisms. This harness-driven design transforms the entire trading pipeline into a sequence of well-defined, reproducible, and auditable decision processes with explicit execution semantics. Extensive experiments on the CSI 300 and S&P 500 benchmarks demonstrate that AlphaCrafter consistently achieves superior risk-adjusted returns while exhibiting substantially lower cross-model and cross-trial variance. These results suggest that harness-based agent design provides a practical foundation for building more reliable, controllable, and robust multi-agent systems for quantitative trading.

cs.AI↗

arXiv2026

Evolving Safety Landscape of Multi-modal Large Language Models: A Survey of Emerging Threats and Safeguards

Multi-modal large language models (MLLMs) integrate heterogeneous modalities through modality alignment and fusion, enabling stronger understanding and reasoning. However, this architectural shift reshapes the safety landscape of machine learning. Increased model complexity and cross-modal interactions give rise to novel threats, including compromised modality integration, modality misalignment, and fused safety risks, reflecting shifts in threat modeling beyond uni-modal assumptions. These shifts, in turn, impose new constraints on safety solutions not captured by existing frameworks rooted in uni-modal learning. Motivated by these challenges, this survey provides a systematic analysis of the evolving safety landscape of MLLMs. We first propose a multimodal grounded taxonomy of safety threats and analyze shifts in threat models, covering adversarial attacks, data poisoning, jailbreaks, and hallucinations. We then summarize updated safety assumptions and organize recent advances in MLLM safety strategies accordingly. Finally, we discuss open challenges and future directions to inform the development of more principled and scalable safety mechanisms for multimodal systems.

cs.LG↗