arXiv subjects

Ying Zhang

Publications and source records attributed to Ying Zhang.

At least 109 records · Page 6Linked to original sources

arXiv2026

Beyond "I cannot fulfill this request": Alleviating Rigid Rejection in LLMs via Label Enhancement

Large Language Models (LLMs) rely on safety alignment to obey safe requests while refusing harmful ones. However, traditional refusal mechanisms often lead to "rigid rejection," where a general template (e.g., "I cannot fulfill this request") indiscriminately triggers refusals and severely undermines the naturalness of interactions between humans and LLMs. To address this issue, LANCE is proposed in this paper to ensure safe yet flexible and natural responses via label enhancement. Specifically, LANCE employs variational inference to perform label enhancement, predicting a continuous distribution across multiple rejection categories. These fine-grained rejection distributions provide multi-way textual gradients for a refinement model to neutralize the hazardous elements in the prompt, so that the LLMs could generate safe responses that avoid rigid rejections while preserving the naturalness of interactions. Experiments demonstrate that LANCE significantly alleviates the rigid rejection problem while maintaining high security standards, significantly outperforming existing baseline models in terms of helpfulness and naturalness of responses.

cs.CL↗

arXiv2026

Robust SGLD algorithm for solving non-convex distributionally robust optimisation problems

In this paper we develop a Stochastic Gradient Langevin Dynamics (SGLD) algorithm tailored for solving a certain class of non-convex distributionally robust optimisation (DRO) problems. By deriving non-asymptotic convergence bounds, we build an algorithm which for any prescribed accuracy $\varepsilon>0$ outputs an estimator whose expected excess risk is at most $\varepsilon$. As a concrete application, we consider the problem of identifying the best non-linear estimator of a given regression model involving a neural network using adversarially corrupted samples. We formulate this problem as a DRO problem and demonstrate both theoretically and numerically the applicability of the proposed robust SGLD algorithm. Moreover, numerical experiments show that the robust SGLD estimator outperforms the estimator obtained using vanilla SGLD in terms of test accuracy, which highlights the advantage of incorporating model uncertainty when optimising with perturbed samples.

math.OC↗

arXiv2026

Relativistic mean-field study of the neutron star inner crust using the asymmetric finite difference method

The ground-state properties of neutron-rich nuclear clusters in the inner crust of neutron stars are investigated within the Wigner-Seitz approximation using a relativistic mean-field framework. The radial Dirac equations are solved with an asymmetric finite-difference scheme, by which the hermiticity is preserved and spurious states are eliminated. Calculations are performed for representative Wigner-Seitz cells employing TM1-based interactions with different symmetry-energy slope parameters $L$, as well as a parametrization with a larger nucleon effective mass. It is found that the binding energy per nucleon decreases systematically with increasing $L$, while a larger effective mass leads to further reduction, particularly at higher densities. Quantum shell effects, which are absent in the Thomas-Fermi approximation, give rise to oscillatory density distributions and modify neutron properties. Within the Wigner-Seitz cell, the resulting neutron root-mean-square radius and chemical potential are shown to be sensitive to both $L$ and the effective nucleon mass, underscoring their important roles in determining the microscopic structure of the neutron-star inner crust.

nucl-th↗

arXiv2026

Generating Proof-of-Vulnerability Tests to Help Enhance the Security of Complex Software

Developers create modern software applications (Apps) on top of third-party libraries (Libs). When library vulnerabilities are reachable through application code, the applications can be vulnerable to software supply chain attacks. Prior work shows that developers often require concrete and executable evidence, i.e., proof-of-vulnerability (PoV) tests, to decide whether a reported dependency vulnerability poses a practical security risk to their application. However, manually crafting such tests is challenging, and existing tool support is insufficient to automate the procedure. To streamline test generation, we created PoVSmith -- a new approach that combines call path analysis, exemplar test, code context, and feedback into multiple prompts to guide a coding agent (i.e., Codex) and a large language model (i.e., GPT) for test generation, execution, and assessment. We evaluated PoVSmith on 33 $\langle$App, Lib$\rangle$ Java program pairs, where each App depends on a vulnerable Lib. PoVSmith revealed 158 unique application-level entry points (i.e., public methods) calling vulnerable library APIs; 152 (96\%) of them were correctly found, together with the call paths properly recognized. With such method call information, PoVSmith generated 152 tests, 84 (55\%) of which demonstrated feasible ways of attacking Apps by exploiting Lib vulnerabilities. PoVSmith substantially outperforms the state-of-the-art LLM-based approach, as it reduces human involvement while dramatically improving test quality. Our work contributes (1) a novel approach of agent-based test generation, (2) an iterative code refinement process driven by execution feedback, and (3) LLM-based quality assessment grounded in both the test context and execution logs.

cs.CR↗

arXiv2026

Deployment-Efficient Short-Term Load Forecasting in AI Data Centers via Sequence-to-Point Knowledge Distillation

Accurately forecasting the bursty and non-stationary power demand of AI data centers has become increasingly important, as abrupt workload-driven variations at the GPU-node level can affect real-time operational efficiency, power management, and grid-data center coordination. However, high-capacity forecasting models are often difficult to deploy at scale because of their memory and latency requirements, while lightweight predictors may fail to capture short-horizon temporal dynamics. To address this accuracy-deployment tradeoff, this paper proposes a deployment-efficient knowledge distillation framework for short-term load forecasting in AI data centers. The proposed framework first trains a high-capacity sequence teacher model for multi-step load trajectory prediction, where residual learning is used to improve robustness under non-stationary operating conditions. A lightweight point-wise student model is then developed for low-latency rolling inference using a compact neural network architecture. To transfer temporal knowledge from the teacher to the student, a sequence-to-point distillation strategy is introduced by aligning near-term predictive behavior and temporally pooled representations. Case studies on the MIT Supercloud dataset demonstrate that the proposed student model improves forecasting accuracy over recent deep learning baselines while reducing the deployment footprint by over 10x in parameter memory and model size.

eess.SY↗

arXiv2026

EDU-CIRCUIT-HW: Evaluating Multimodal Large Language Models on Real-World University-Level STEM Student Handwritten Solutions

Multimodal Large Language Models (MLLMs) hold significant promise for revolutionizing traditional education and reducing teachers' workload. However, accurately interpreting unconstrained STEM student handwritten solutions with intertwined mathematical formulas, diagrams, and textual reasoning poses a significant challenge due to the lack of authentic and domain-specific benchmarks. Additionally, current evaluation paradigms predominantly rely on the outcomes of downstream tasks (e.g., auto-grading), which often probe only a subset of the recognized content, thereby failing to capture the MLLMs' understanding of complex handwritten logic as a whole. To bridge this gap, we release EDU-CIRCUIT-HW, a dataset consisting of 1,300+ authentic student handwritten solutions from a university-level STEM course. Utilizing the expert-verified verbatim transcriptions and grading reports of student solutions, we simultaneously evaluate various MLLMs' upstream recognition fidelity and downstream auto-grading performance. Our evaluation uncovers an astonishing scale of latent failures within MLLM-recognized student handwritten content, highlighting the models' insufficient reliability for auto-grading and other understanding-oriented applications in high-stakes educational settings. As a potential solution, we present a case study demonstrating that leveraging identified error patterns to preemptively detect and correct recognition errors, while requiring only minimal human intervention (e.g., routing 3.3% of assignments to human graders and the remainder to the GPT-5.1 grader), can effectively enhance the robustness of the deployed AI-enabled grading system. Code and dataset are available in this GitHub repo: https://gt-learning-innovation.github.io/CIRCUIT_EDU_HW_ACL.

cs.CV↗

arXiv2026

Multiresonant Membrane Metasurfaces for Multifunctional Fingerprint Recognition and Real-time Biochemical Tracking

Label-free identification and real-time tracking of biochemical substances became critical for molecular diagnostics and chemical analysis, yet conventional resonant terahertz metasurface sensing relies on a single resonance, limiting spectral selectivity and dynamic capability. Here, we suggest multiresonant membrane metasurfaces and implement them for simultaneous static molecular fingerprint retrieval and dynamic reaction monitoring within a single pixel. We consider a membrane metasurface supporting multiple quasi-bound states in the continuum designed at target frequencies and enabling the tailoring of the field enhancement and frequency-selective interaction with target analytes. As a proof-of-concept, we achieve label-free detection of the dual fingerprint absorption features of pefloxacin at 0.78 THz and 0.99 THz, and real-time tracking of vitamin C oxidation and denaturation under ambient conditions. The kinetic profiles extracted from the THz amplitude evolution show excellent agreement with nonlinear reaction models, demonstrating quantitative biochemical tracking capabilities. Our results establish a versatile and scalable THz photonic platform that unifies static fingerprint identification and dynamic reaction monitoring, paving the way toward integrated on-chip biochemical analytics and multifunctional metasurface sensors.

physics.optics↗

arXiv2026

Regime-Adaptive Weighted Ensemble Learning for Computing-Driven Dynamic Load Forecasting in AI Data Centers

Short-term load forecasting for AI data centers presents new challenges because it is computing-driven, with heterogeneous job arrivals, sizes, and durations exhibiting bursty, non-stationary dynamics. Compared with traditional load types, data center loads are less researched and can pose greater threats to the efficiency and stability of power grids. To close the gap, this paper proposes a regime-adaptive ensemble learning forecasting algorithm to predict computing-driven dynamic workloads in AI data centers. A weight-learned neural network within an ensemble learning framework is developed to exploit the complementary strengths of two machine learning (ML) submodels across varying operating regimes. Furthermore, a novel feature engineering strategy is developed to incrementally learn from a non-stationary data stream. Thus, the ensemble weights are dynamically optimized to facilitate adaptive calibration of inter-submodel contributions. Comparative case studies on the MIT Supercloud dataset demonstrate that the proposed method significantly enhances load forecasting accuracy and adaptivity across various regimes, and the selected combination of ML models for ensemble learning outperforms other possible combinations. To the best of our knowledge, our method is the first to reduce minute-class forecasting errors for AI data center loads to below 1%, highlighting its potential for grid-interactive coordination and demand response.

eess.SY↗

arXiv2026

Strong convergence and temporal-spatial regularity for tamed Euler approximations of Lévy-driven SDEs

We study the temporal-spatial regularity properties of tamed Euler approximations for Lévy-driven SDEs with superlinearly growing drift and diffusion coefficients. We first introduce a novel tamed Euler-type scheme and establish its strong convergence. We then derive temporal-spatial regularity estimates with respect to the initial value, the initial time, and the evaluation time. In particular, we obtain a stability estimate for fixed step size and a corresponding continuity estimate in the vanishing step-size regime. Numerical experiments are presented to support the theoretical results.

math.NA↗

arXiv2026

ESICA: A Scalable Framework for Text-Guided 3D Medical Image Segmentation

Text guided 3D medical image segmentation offers a flexible alternative to class based and spatial prompt based models by allowing users to specify regions of interest directly in natural language. This paradigm avoids reliance on predefined label sets, reduces ambiguous outputs, and aligns more naturally with clinical workflows. However, existing text guided frameworks are often computationally expensive, exhibit weak text volume feature alignment, and fail to capture fine anatomical details. We propose ESICA, a lightweight and scalable framework that addresses these challenges through three innovations: (1) a similarity matrix based mask prediction formulation that enhances semantic alignment, (2) an efficient decomposed decoder with adapter modules for accurate volumetric decoding, and (3) a two pass refinement strategy that sharpens boundaries and resolves uncertain regions. To improve training stability and generalization, ESICA adopts a two stage scheme consisting of positive only pretraining followed by balanced fine tuning. On the CVPR BiomedSegFM benchmark spanning five imaging modalities (CT, MRI, PET, ultrasound, and microscopy), ESICA achieves state of the art segmentation accuracy, while the compact ESICA4 Lite variant attains similar segmentation performance with substantially fewer parameters, yielding a superior efficiency accuracy trade off. Our framework advances text guided segmentation toward efficient, scalable, and clinically deployable systems. Code will be made publicly available at https://github.com/mirthAI/ESICA.

cs.CV↗

arXiv2026

Physics-Aware LLM-Based Probabilistic Wind Power Scenario Generation under Extreme Icing Conditions

Accurately characterizing wind power uncertainty under icing and post-disaster conditions remains a critical challenge for resilient power system operation. To address this issue, this paper proposes a physics-aware large language model (LLM) framework for probabilistic wind power scenario generation under extreme icing conditions. The proposed framework integrates supervisory control and data acquisition (SCADA)-based physical modeling, multimodal tokenization, and a causal Transformer architecture trained in an autoregressive manner. A physics-aware decoding scheme effectively enforces rated power limits and ramping constraints on the generated trajectories while preserving stochastic diversity. Case studies using real wind turbine data show that the proposed method reproduces icing-induced power degradation and temporal variability observed during extreme weather. The resulting scenarios are physically consistent and high-fidelity, thereby significantly enhancing resilience assessment and recovery planning in renewable-integrated power systems.

eess.SY↗

arXiv2026

Bayesian Inference of Dense-Matter Equations of State from Small-Radius Compact Stars with Twin-Star Scenarios

We investigate dense-matter equations of state (EOSs) within a Bayesian framework, with particular emphasis on whether recent small-radius compact-star candidates can be accommodated in a twin-star scenario. For the hadronic sector, we adopt a meta-modeling EOS constrained by the NICER mass--radius measurements of PSR J0030$+$0451, PSR J0437$-$4715, PSR J0614$-$3329, and the massive pulsar PSR J0740$+$6620. The hadronic inference indicates that PSR J0614$-$3329 favors a somewhat softer EOS than the other two $\sim1.4\,M_\odot$ pulsars, while the $\sim2\,M_\odot$ constraint prevents the EOS from becoming too soft. We then introduce a strong first-order phase transition through a constant-speed-of-sound quark-matter segment. Using HESS J1731$-$347 and XTE J1814$-$338 to constrain the phase-transition parameters, we find a preferred transition density of $n_\mathrm{t}\sim2.7\text{--}2.8\,n_0$, a sizable energy-density jump of $600\text{--}700$ MeV, and a relatively large post-transition sound speed of $c_s^2/c^2\sim0.85$. Such a phase transition generates a disconnected hybrid branch with radii of about $6\text{--}7$ km at masses around $1.2\text{--}1.4\,M_\odot$, and strongly suppresses the dimensionless tidal deformability relative to the purely hadronic branch. This pronounced change in tidal deformability is a characteristic signature of the twin-star mechanism and may provide an important observational tool for identifying phase transitions in neutron-star matter in future multimessenger measurements. These results show that small-radius compact stars can provide direct constraints on both the strength of a first-order phase transition and the stiffness of the post-transition phase in dense matter.

astro-ph.HE↗

arXiv2026

Multi-Perspective Evidence Synthesis and Reasoning for Unsupervised Multimodal Entity Linking

Multimodal Entity Linking (MEL) is a fundamental task in data management that maps ambiguous mentions with diverse modalities to the multimodal entities in a knowledge base. However, most existing MEL approaches primarily focus on optimizing instance-centric features and evidence, leaving broader forms of evidence and their intricate interdependencies insufficiently explored. Motivated by the observation that human expert decision-making process relies on multi-perspective judgment, in this work, we propose MSR-MEL, a Multi-perspective Evidence Synthesis and Reasoning framework with Large Language Models (LLMs) for unsupervised MEL. Specifically, we adopt a two-stage framework: (1) Offline Multi-Perspective Evidence Synthesis constructs a comprehensive set of evidence. This includes instance-centric evidence capturing the instance-centric multimodal information of mentions and entities, group-level evidence that aggregates neighborhood information, lexical evidence based on string overlap ratio, and statistical evidence based on simple summary statistics. A core contribution of our framework is the synthesis of group-level evidence, which effectively aggregates vital neighborhood information by graph. We first construct LLM-enhanced contextualized graphs. Subsequently, different modalities are jointly aligned through an asymmetric teacher-student graph neural network. (2) Online Multi-Perspective Evidence Reasoning leverages the power of LLM as a reasoning module to analyze the correlation and semantics of the multi-perspective evidence to induce an effective ranking strategy for accurate entity linking without supervision. Extensive experiments on widely used MEL benchmarks demonstrate that MSR-MEL consistently outperforms state-of-the-art unsupervised methods. The source code of this paper was available at: https://anonymous.4open.science/r/MSR-MEL-C21E/.

cs.CL↗

arXiv2026

Cultural Newcomers Dining Across Borders: Need-Based Design Envision of Mixed Media Integration in MR for Foreign Menu Understanding and Ordering

Cultural newcomers (CNs), including new immigrants and international students, often encounter cognitive barriers and social anxiety, exacerbated by unfamiliar cultural terminology in daily interactions. This research examines these challenges in the context of ordering in foreign restaurants. Current translation tools have significant limitations in their information delivery with current media presentation methods. This research investigates the challenges and needs of CNs in ordering scenarios in a foreign restaurant through interview sessions (N = 13) and explored their expectation of mixed media integration (Image, Video, 3D Model) through a participatory design session that featured an immersive restaurant experience to support brainstorming. Based on qualitative analysis of participants' needs and expectations, the mixed media ordering assistant is conceptualized across 4 key dimensions: Key Features, User interaction, Media hierarchy, and Information presentation, with the objective of alleviating cultural barrier, linguistic barrier, cognitive load and improving the dining experience for CNs.

cs.HC↗

arXiv2026

Structure-Adaptive Sparse Diffusion in Voxel Space for 3D Medical Image Enhancement

Three-dimensional (3D) medical image enhancement, including denoising and super-resolution, is critical for clinical diagnosis in CT, PET, and MRI. Although diffusion models have shown remarkable success in 2D medical imaging, scaling them to high-resolution 3D volumes remains computationally prohibitive due to lengthy diffusion trajectories over high-dimensional volumetric data. We observe that in conditional enhancement, strong anatomical priors in the degraded input render dense noise schedules largely redundant. Leveraging this insight, we propose a sparse voxel-space diffusion framework that trains and samples on a compact set of uniformly subsampled timesteps. The network predicts clean data directly on the data manifold, supervised in velocity space for stable gradient scaling. A lightweight Structure-aware Trajectory Modulation (STM) module recalibrates time embeddings at each network block based on local anatomical content, enabling structure-adaptive denoising over the shared sparse schedule. Operating directly in voxel space, our framework preserves fine anatomical detail without lossy compression while achieving up to $10\times$ training acceleration. Experiments on four datasets spanning CT, PET, and MRI demonstrate state-of-the-art performance on both denoising and super-resolution tasks. Our code is publicly available at: https://github.com/mirthAI/sparse-3d-diffusion.

cs.CV↗

arXiv2026

IF-CRITIC: Towards a Fine-Grained LLM Critic for Instruction-Following Evaluation

Instruction-following is a fundamental ability of Large Language Models (LLMs), requiring their generated outputs to follow multiple constraints imposed in input instructions. Numerous studies have attempted to enhance this ability through preference optimization or reinforcement learning based on reward signals from LLM-as-a-Judge. However, existing evaluation models for instruction-following still possess many deficiencies, such as substantial costs and unreliable assessments. To this end, we propose IF-CRITIC, an LLM critic for fine-grained, efficient, and reliable instruction-following evaluation. We first develop a checklist generator to decompose instructions and generate constraint checklists. With the assistance of the checklists, we collect high-quality critique training data through a multi-stage critique filtering mechanism and employ a constraint-level preference optimization method to train IF-CRITIC. Extensive experiments show that the evaluation performance of IF-CRITIC can beat strong LLM-as-a-Judge baselines, including o4-mini and Gemini-3-Pro. With the reward signals provided by IF-CRITIC, LLMs can achieve substantial performance gains in instruction-following optimization under lower computational overhead compared to strong LLM critic baselines. Our code and model are available at https://github.com/thu-coai/IF-CRITIC.

cs.CL↗

arXiv2026

IF-RewardBench: Benchmarking Judge Models for Instruction-Following Evaluation

Instruction-following is a foundational capability of large language models (LLMs), with its improvement hinging on scalable and accurate feedback from judge models. However, the reliability of current judge models in instruction-following remains underexplored due to several deficiencies of existing meta-evaluation benchmarks, such as their insufficient data coverage and oversimplified pairwise evaluation paradigms that misalign with model optimization scenarios. To this end, we propose IF-RewardBench, a comprehensive meta-evaluation benchmark for instruction-following that covers diverse instruction and constraint types. For each instruction, we construct a preference graph containing all pairwise preferences among multiple responses based on instruction-following quality. This design enables a listwise evaluation paradigm that assesses the capabilities of judge models to rank multiple responses, which is essential in guiding model alignment. Extensive experiments on IF-RewardBench reveal significant deficiencies in current judge models and demonstrate that our benchmark achieves a stronger positive correlation with downstream task performance compared to existing benchmarks. Our codes and data are available at https://github.com/thu-coai/IF-RewardBench.

cs.CL↗

arXiv2026

Seek-and-Solve: Benchmarking MLLMs for Visual Clue-Driven Reasoning in Daily Scenarios

Daily scenarios are characterized by visual richness, requiring Multimodal Large Language Models (MLLMs) to filter noise and identify decisive visual clues for accurate reasoning. Yet, current benchmarks predominantly aim at evaluating MLLMs' pre-existing knowledge or perceptual understanding, often neglecting the critical capability of reasoning. To bridge this gap, we introduce DailyClue, a benchmark designed for visual clue-driven reasoning in daily scenarios. Our construction is guided by two core principles: (1) strict grounding in authentic daily activities, and (2) challenging query design that necessitates more than surface-level perception. Instead of simple recognition, our questions compel MLLMs to actively explore suitable visual clues and leverage them for subsequent reasoning. To this end, we curate a comprehensive dataset spanning four major daily domains and 16 distinct subtasks. Comprehensive evaluation across MLLMs and agentic models underscores the formidable challenge posed by our benchmark. Our analysis reveals several critical insights, emphasizing that the accurate identification of visual clues is essential for robust reasoning.

cs.CV↗