arXiv subjects

Shuo Wang

Publications and source records attributed to Shuo Wang.

At least 181 records · Page 10Linked to original sources

arXiv2026

Teaching LLMs to Learn Tool Trialing and Execution through Environment Interaction

Equipping Large Language Models (LLMs) with external tools enables them to solve complex real-world problems. However, the robustness of existing methods remains a critical challenge when confronting novel or evolving tools. Existing trajectory-centric paradigms primarily rely on memorizing static solution paths during training, which limits the ability of LLMs to generalize tool usage to newly introduced or previously unseen tools. In this paper, we propose ToolMaster, a framework that shifts tool use from imitating golden tool-calling trajectories to actively learning tool usage through interaction with the environment. To optimize LLMs for tool planning and invocation, ToolMaster adopts a trial-and-execution paradigm, which trains LLMs to first imitate teacher-generated trajectories containing explicit tool trials and self-correction, followed by reinforcement learning to coordinate the trial and execution phases jointly. This process enables agents to autonomously explore correct tool usage by actively interacting with environments and forming experiential knowledge that benefits tool execution. Experimental results demonstrate that ToolMaster significantly outperforms existing baselines in terms of generalization and robustness across unseen or unfamiliar tools. All code and data are available at https://github.com/NEUIR/ToolMaster.

cs.SE↗

arXiv2026

Long-Chain Reasoning Distillation via Adaptive Prefix Alignment

Large Language Models (LLMs) have demonstrated remarkable reasoning capabilities, particularly in solving complex mathematical problems. Recent studies show that distilling long reasoning trajectories can effectively enhance the reasoning performance of small-scale student models. However, teacher-generated reasoning trajectories are often excessively long and structurally complex, making them difficult for student models to learn. This mismatch leads to a gap between the provided supervision signal and the learning capacity of the student model. To address this challenge, we propose Prefix-ALIGNment distillation (P-ALIGN), a framework that fully exploits teacher CoTs for distillation through adaptive prefix alignment. Specifically, P-ALIGN adaptively truncates teacher-generated reasoning trajectories by determining whether the remaining suffix is concise and sufficient to guide the student model. Then, P-ALIGN leverages the teacher-generated prefix to supervise the student model, encouraging effective prefix alignment. Experiments on multiple mathematical reasoning benchmarks demonstrate that P-ALIGN outperforms all baselines by over 3%. Further analysis indicates that the prefixes constructed by P-ALIGN provide more effective supervision signals, while avoiding the negative impact of redundant and uncertain reasoning components. All code is available at https://github.com/NEUIR/P-ALIGN.

cs.CL↗

arXiv2026

GI-Bench: A Panoramic Benchmark Revealing the Knowledge-Experience Dissociation of Multimodal Large Language Models in Gastrointestinal Endoscopy Against Clinical Standards

Multimodal Large Language Models (MLLMs) show promise in gastroenterology, yet their performance against comprehensive clinical workflows and human benchmarks remains unverified. To systematically evaluate state-of-the-art MLLMs across a panoramic gastrointestinal endoscopy workflow and determine their clinical utility compared with human endoscopists. We constructed GI-Bench, a benchmark encompassing 20 fine-grained lesion categories. Twelve MLLMs were evaluated across a five-stage clinical workflow: anatomical localization, lesion identification, diagnosis, findings description, and management. Model performance was benchmarked against three junior endoscopists and three residency trainees using Macro-F1, mean Intersection-over-Union (mIoU), and multi-dimensional Likert scale. Gemini-3-Pro achieved state-of-the-art performance. In diagnostic reasoning, top-tier models (Macro-F1 0.641) outperformed trainees (0.492) and rivaled junior endoscopists (0.727; p>0.05). However, a critical "spatial grounding bottleneck" persisted; human lesion localization (mIoU >0.506) significantly outperformed the best model (0.345; p<0.05). Furthermore, qualitative analysis revealed a "fluency-accuracy paradox": models generated reports with superior linguistic readability compared with humans (p<0.05) but exhibited significantly lower factual correctness (p<0.05) due to "over-interpretation" and hallucination of visual features. GI-Bench maintains a dynamic leaderboard that tracks the evolving performance of MLLMs in clinical endoscopy. The current rankings and benchmark results are available at https://roterdl.github.io/GIBench/.

cs.CV↗

arXiv2026

Revealing the Attention Floating Mechanism in Masked Diffusion Models

Masked diffusion models (MDMs), which leverage bidirectional attention and a denoising process, are narrowing the performance gap with autoregressive models (ARMs). However, their internal attention mechanisms remain under-explored. This paper investigates the attention behaviors in MDMs, revealing the phenomenon of Attention Floating. Unlike ARMs, where attention converges to a fixed sink, MDMs exhibit dynamic, dispersed attention anchors that shift across denoising steps and layers. Further analysis reveals its Shallow Structure-Aware, Deep Content-Focused attention mechanism: shallow layers utilize floating tokens to build a global structural framework, while deeper layers allocate more capability toward capturing semantic content. Empirically, this distinctive attention pattern provides a mechanistic explanation for the strong in-context learning capabilities of MDMs, allowing them to double the performance compared to ARMs in knowledge-intensive tasks. All codes and datasets are available at https://github.com/NEUIR/Attention-Floating.

cs.LG↗

arXiv2026

Empirical Analysis of Decoding Biases in Masked Diffusion Models

cs.AI↗

arXiv2026

Punctuation-aware Hybrid Trainable Sparse Attention for Large Language Models

Attention serves as the fundamental mechanism for long-context modeling in large language models (LLMs), yet dense attention becomes structurally prohibitive for long sequences due to its quadratic complexity. Consequently, sparse attention has received increasing attention as a scalable alternative. However, existing sparse attention methods rely on coarse-grained semantic representations during block selection, which blur intra-block semantic boundaries and lead to the loss of critical information. To address this issue, we propose \textbf{P}unctuation-aware \textbf{H}ybrid \textbf{S}parse \textbf{A}ttention \textbf{(PHSA)}, a natively trainable sparse attention framework that leverages punctuation tokens as semantic boundary anchors. Specifically, (1) we design a dual-branch aggregation mechanism that fuses global semantic representations with punctuation-enhanced boundary features, preserving the core semantic structure while introducing almost no additional computational overhead; (2) we introduce an extreme-sparsity-adaptive training and inference strategy that stabilizes model behavior under very low token activation ratios; Extensive experiments on general benchmarks and long-context evaluations demonstrate that PHSA consistently outperforms dense attention and state-of-the-art sparse attention baselines, including InfLLM v2. Specifically, for the 0.6B-parameter model with 32k-token input sequences, PHSA can reduce the information loss by 10.8\% at a sparsity ratio of 97.3\%.

cs.CL↗

arXiv2026

A Machine Learning Framework for Off Ball Defensive Role and Performance Evaluation in Football

Evaluating off-ball defensive performance in football is challenging, as traditional metrics do not capture the nuanced coordinated movements that limit opponent action selection and success probabilities. Although widely used possession value models excel at appraising on-ball actions, their application to defense remains limited. Existing counterfactual methods, such as ghosting models, help extend these analyses but often rely on simulating "average" behavior that lacks tactical context. To address this, we introduce a covariate-dependent Hidden Markov Model (CDHMM) tailored to corner kicks, a highly structured aspect of football games. Our label-free model infers time-resolved man-marking and zonal assignments directly from player tracking data. We leverage these assignments to propose a novel framework for defensive credit attribution and a role-conditioned ghosting method for counterfactual analysis of off-ball defensive performance. We show how these contributions provide a interpretable evaluation of defensive contributions against context-aware baselines.

cs.LG↗

arXiv2025

One-shot synthesis of rare gastrointestinal lesions improves diagnostic accuracy and clinical training

Rare gastrointestinal lesions are infrequently encountered in routine endoscopy, restricting the data available for developing reliable artificial intelligence (AI) models and training novice clinicians. Here we present EndoRare, a one-shot, retraining-free generative framework that synthesizes diverse, high-fidelity lesion exemplars from a single reference image. By leveraging language-guided concept disentanglement, EndoRare separates pathognomonic lesion features from non-diagnostic attributes, encoding the former into a learnable prototype embedding while varying the latter to ensure diversity. We validated the framework across four rare pathologies (calcifying fibrous tumor, juvenile polyposis syndrome, familial adenomatous polyposis, and Peutz-Jeghers syndrome). Synthetic images were judged clinically plausible by experts and, when used for data augmentation, significantly enhanced downstream AI classifiers, improving the true positive rate at low false-positive rates. Crucially, a blinded reader study demonstrated that novice endoscopists exposed to EndoRare-generated cases achieved a 0.400 increase in recall and a 0.267 increase in precision. These results establish a practical, data-efficient pathway to bridge the rare-disease gap in both computer-aided diagnostics and clinical education.

cs.CV↗

arXiv2025

Entropy-Guided Token Dropout: Training Autoregressive Language Models with Limited Domain Data

As access to high-quality, domain-specific data grows increasingly scarce, multi-epoch training has become a practical strategy for adapting large language models (LLMs). However, autoregressive models often suffer from performance degradation under repeated data exposure, where overfitting leads to a marked decline in model capability. Through empirical analysis, we trace this degradation to an imbalance in learning dynamics: predictable, low-entropy tokens are learned quickly and come to dominate optimization, while the model's ability to generalize on high-entropy tokens deteriorates with continued training. To address this, we introduce EntroDrop, an entropy-guided token dropout method that functions as structured data regularization. EntroDrop selectively masks low-entropy tokens during training and employs a curriculum schedule to adjust regularization strength in alignment with training progress. Experiments across model scales from 0.6B to 8B parameters show that EntroDrop consistently outperforms standard regularization baselines and maintains robust performance throughout extended multi-epoch training. These findings underscore the importance of aligning regularization with token-level learning dynamics when training on limited data. Our approach offers a promising pathway toward more effective adaptation of LLMs in data-constrained domains.

cs.CL↗

arXiv2025

Robust Polyp Detection and Diagnosis through Compositional Prompt-Guided Diffusion Models

Colorectal cancer (CRC) is a significant global health concern, and early detection through screening plays a critical role in reducing mortality. While deep learning models have shown promise in improving polyp detection, classification, and segmentation, their generalization across diverse clinical environments, particularly with out-of-distribution (OOD) data, remains a challenge. Multi-center datasets like PolypGen have been developed to address these issues, but their collection is costly and time-consuming. Traditional data augmentation techniques provide limited variability, failing to capture the complexity of medical images. Diffusion models have emerged as a promising solution for generating synthetic polyp images, but the image generation process in current models mainly relies on segmentation masks as the condition, limiting their ability to capture the full clinical context. To overcome these limitations, we propose a Progressive Spectrum Diffusion Model (PSDM) that integrates diverse clinical annotations-such as segmentation masks, bounding boxes, and colonoscopy reports-by transforming them into compositional prompts. These prompts are organized into coarse and fine components, allowing the model to capture both broad spatial structures and fine details, generating clinically accurate synthetic images. By augmenting training data with PSDM-generated samples, our model significantly improves polyp detection, classification, and segmentation. For instance, on the PolypGen dataset, PSDM increases the F1 score by 2.12% and the mean average precision by 3.09%, demonstrating superior performance in OOD scenarios and enhanced generalization.

cs.CV↗

arXiv2025

C$^2$GSPG: Confidence-calibrated Group Sequence Policy Gradient towards Self-aware Reasoning

Reinforcement Learning (RL) methods, exemplified by Group Relative Policy Optimization (GRPO) and its variants, play a central role in developing reasoning models. However, these methods often suffer from a critical overconfidence issue, which prevents them from achieving self-aware reasoning models. In this study, we propose a simple yet effective confidence-calibration group sequence policy gradient method, called C$^2$GSPG, which simultaneously enhances reasoning performance while suppressing overconfidence. In principle, we propose a Group Sequence Policy Gradient (GSPG) framework for learning reasoning models, which eliminates the token-level bias commonly appearing in GRPO and its variants. In this framework, we define the model confidence for each reasoning problem using the normalized sequence-level probability, and then apply a cross-entropy regularizer to calibrate the model confidence to the sequence's reward. We demonstrate that the confidence calibration regularizer and GSPG are collaborative for binary rewards, as their objectives always share the same gradient direction. For non-binary rewards, we apply nonlinear reward normalization and adaptive regularizer clipping, mitigating the potential conflict between the two objectives. Applying C$^2$GSPG to post-train large language models in logical and mathematical reasoning tasks, we show its superiority over state-of-the-art methods in both reasoning accuracy and confidence calibration. The code of C$^2$GSPG is available at https://github.com/HaotianLiu123/CCGSPG.

cs.LG↗

arXiv2025

Magnetically tunable singularity of Moire bound states in the continuum

We investigate the band structure and the dynamics of topological singularities of magneto-optical Moiré bound states in the continuum (BICs), which originate from two overlapped conventional magneto-optical BIC gratings with mismatched periods. We show that breaking time-reversal symmetry by applying an external magnetic field in different directions can effectively manipulate the Moiré BICs' topological singularities, which exhibit a high Q factor throughout the momentum space.

physics.optics↗

arXiv2025

ExpShield: Safeguarding Web Text from Unauthorized Crawling and LLM Exploitation

As large language models increasingly memorize web-scraped training content, they risk exposing copyrighted or private information. Existing protections require compliance from crawlers or model developers, fundamentally limiting their effectiveness. We propose ExpShield, a proactive self-guard that mitigates memorization while maintaining readability via invisible perturbations, and we formulate it as a constrained optimization problem. Due to the lack of an individual-level risk metric for natural text, we first propose instance exploitation, a metric that measures how much training on a specific text increases the chance of guessing that text from a set of candidates-with zero indicating perfect defense. Directly solving the problem is infeasible for defenders without sufficient knowledge, thus we develop two effective proxy solutions: single-level optimization and synthetic perturbation. To enhance the defense, we reveal and verify the memorization trigger hypothesis, which can help to identify key tokens for memorization. Leveraging this insight, we design targeted perturbations that (i) neutralize inherent trigger tokens to reduce memorization and (ii) introduce artificial trigger tokens to misdirect model memorization. Experiments validate our defense across attacks, model scales, and tasks in language and vision-to-language modeling. Even with privacy backdoor, the Membership Inference Attack (MIA) AUC drops from 0.95 to 0.55 under the defense, and the instance exploitation approaches zero. This suggests that compared to the ideal no-misuse scenario, the risk of exposing a text instance remains nearly unchanged despite its inclusion in the training data.

cs.CR↗

arXiv2025

Switchable half-quantum flux states in a ring of the kagome superconductor CsV$_3$Sb$_5$

Magnetic flux quantization in units of $Φ_0 = h/2e$ is a defining feature of superconductivity, rooted in the charge-2e nature of Cooper pairs. In a ring geometry, the flux quantization leads to oscillations in the critical temperature with magnetic flux, known as the Little-Parks effect. While the maximal critical temperature is conventionally at zero flux, departures from this rule, for instance shifts by a half-quantum flux $Φ_0/2$, clearly signal unconventional superconducting states and require sign-changing order parameters. Historically, such $π$-phase shifts in Little-Parks oscillations have been found in tricrystals or engineered ring structures that intentionally incorporate a $π$-phase shift. Here we report the discovery of switchable half-quantum flux states in rings made from single crystals of the kagome superconductor CsV$_3$Sb$_5$. We observe Little-Parks oscillations with a $π$-phase shift at zero bias current, which can be reversibly tuned to conventional Little-Parks oscillations upon applying a bias current. Between the $π$-phase and 0-phase regimes, $h/4e$ periodic oscillations appear. Our observations suggest unconventional pairing, potentially in the form of a multicomponent order parameter in the kagome superconductor CsV$_3$Sb$_5$, and reveal an electrically tunable landscape of competing superconducting condensates and fractional flux states.

cond-mat.supr-con↗

arXiv2025

RoleRMBench & RoleRM: Towards Reward Modeling for Profile-Based Role Play in Dialogue Systems

Reward modeling has become a cornerstone of aligning large language models (LLMs) with human preferences. Yet, when extended to subjective and open-ended domains such as role play, existing reward models exhibit severe degradation, struggling to capture nuanced and persona-grounded human judgments. To address this gap, we introduce RoleRMBench, the first systematic benchmark for reward modeling in role-playing dialogue, covering seven fine-grained capabilities from narrative management to role consistency and engagement. Evaluation on RoleRMBench reveals large and consistent gaps between general-purpose reward models and human judgment, particularly in narrative and stylistic dimensions. We further propose RoleRM, a reward model trained with Continuous Implicit Preferences (CIP), which reformulates subjective evaluation as continuous consistent pairwise supervision under multiple structuring strategies. Comprehensive experiments show that RoleRM surpasses strong open- and closed-source reward models by over 24% on average, demonstrating substantial gains in narrative coherence and stylistic fidelity. Our findings highlight the importance of continuous preference representation and annotation consistency, establishing a foundation for subjective alignment in human-centered dialogue systems.

cs.CL↗

arXiv2025

UPETrack: Unidirectional Position Estimation for Tracking Occluded Deformable Linear Objects

Real-time state tracking of Deformable Linear Objects (DLOs) is critical for enabling robotic manipulation of DLOs in industrial assembly, medical procedures, and daily-life applications. However, the high-dimensional configuration space, nonlinear dynamics, and frequent partial occlusions present fundamental barriers to robust real-time DLO tracking. To address these limitations, this study introduces UPETrack, a geometry-driven framework based on Unidirectional Position Estimation (UPE), which facilitates tracking without the requirement for physical modeling, virtual simulation, or visual markers. The framework operates in two phases: (1) visible segment tracking is based on a Gaussian Mixture Model (GMM) fitted via the Expectation Maximization (EM) algorithm, and (2) occlusion region prediction employing UPE algorithm we proposed. UPE leverages the geometric continuity inherent in DLO shapes and their temporal evolution patterns to derive a closed-form positional estimator through three principal mechanisms: (i) local linear combination displacement term, (ii) proximal linear constraint term, and (iii) historical curvature term. This analytical formulation allows efficient and stable estimation of occluded nodes through explicit linear combinations of geometric components, eliminating the need for additional iterative optimization. Experimental results demonstrate that UPETrack surpasses two state-of-the-art tracking algorithms, including TrackDLO and CDCPD2, in both positioning accuracy and computational efficiency.

cs.RO↗

arXiv2025

A study of the spectrum resource leasing method based on ERC4907 extension

The ERC4907 standard enables rentable Non-Fungible Tokens (NFTs) but is limited to single-user, single-time-slot authorization, which severely limits its applicability and efficiency in decentralized multi-slot scheduling scenarios. To address this limitation, this paper proposes Multi-slot ERC4907 (M-ERC4907) extension method. The M-ERC4907 method introduces novel functionalities to support the batch configuration of multiple time slots and simultaneous authorization of multiple users, thereby effectively eliminating the rigid sequential authorization constraint of ERC4907. The experiment was conducted on the Remix development platform. Experimental results show that the M-ERC4907 method significantly reduces on-chain transactions and overall Gas consumption, leading to enhanced scalability and resource allocation efficiency.

cs.DC↗

arXiv2025

Towards Modality- and Sampling-Universal Learning Strategies for Accelerating Cardiovascular Imaging: Summary of the CMRxRecon2024 Challenge

Cardiovascular health is vital to human well-being, and cardiac magnetic resonance (CMR) imaging is considered the {clinical reference standard} for diagnosing cardiovascular disease. However, its adoption is hindered by long scan times, complex contrasts, and inconsistent quality. While deep learning methods perform well on specific CMR imaging {sequences}, they often fail to generalize across modalities and sampling schemes. The lack of benchmarks for high-quality, fast CMR image reconstruction further limits technology comparison and adoption. The CMRxRecon2024 challenge, attracting over 200 teams from 18 countries, addressed these issues with two tasks: generalization to unseen {modalities} and robustness to diverse undersampling patterns. We introduced the largest public multi-{modality} CMR raw dataset, an open benchmarking platform, and shared code. Analysis of the best-performing solutions revealed that prompt-based adaptation and enhanced physics-driven consistency enabled strong cross-scenario performance. These findings establish principles for generalizable reconstruction models and advance clinically translatable AI in cardiovascular imaging.

eess.IV↗