arXiv subjects

Zelin Cao

Publications and source records attributed to Zelin Cao.

4 recordsLinked to original sources

arXiv2026

Structure-BiEval: A Self-Supervised, Dual-Track Framework for Decoupling Structure and Content in LLM Evaluation for Web Information Systems

As Large Language Models (LLMs) evolve into the core of Web-based autonomous agents and complex Web Information Systems, their ability to faithfully translate natural language into rigorous structured formats has become paramount, as this capability is critical for Web API invocation and data exchange. However, evaluating this structural fidelity in Web-native payloads remains a challenge: traditional text metrics fail to capture topological consistency in semi-structured Web data, while manual evaluation is prohibitively costly. To address this, we propose Structure-BiEval, a novel self-supervised framework for quantitative, annotation-free assessment tailored for Web data engineering. By leveraging deterministic Intermediate Representations, our framework effectively decouples structure from content, utilizing Content Semantic Accuracy and Normalized Tree Edit Distance as precise metrics. We empirically benchmark 15 state-of-the-art LLMs across dual Web structural topologies, namely Hierarchical Data (Web backend payloads) and Tabular Data (Web frontend presentation). The results reveal substantial variability in structural performance, including cases where mid-sized models unexpectedly outperform larger counterparts in Web data formatting. Furthermore, our findings show that deep recursive nesting poses a consistent challenge for Web agents across varying parameter scales.

cs.CL↗

arXiv2024

Explicitly Integrating Judgment Prediction with Legal Document Retrieval: A Law-Guided Generative Approach

Legal document retrieval and judgment prediction are crucial tasks in intelligent legal systems. In practice, determining whether two documents share the same judgments is essential for establishing their relevance in legal retrieval. However, existing legal retrieval studies either ignore the vital role of judgment prediction or rely on implicit training objectives, expecting a proper alignment of legal documents in vector space based on their judgments. Neither approach provides explicit evidence of judgment consistency for relevance modeling, leading to inaccuracies and a lack of transparency in retrieval. To address this issue, we propose a law-guided method, namely GEAR, within the generative retrieval framework. GEAR explicitly integrates judgment prediction with legal document retrieval in a sequence-to-sequence manner. Experiments on two Chinese legal case retrieval datasets show the superiority of GEAR over state-of-the-art methods while maintaining competitive judgment prediction performance. Moreover, we validate its robustness across languages and domains on a French statutory article retrieval dataset.

cs.IR↗

arXiv2019

Transmutation elements Re/Ta effect on vacancy formation and dissociation behavior in W

Transmutation elements are the essential products in plasma facing materials tungsten, and have further effects on point defects evolution resulted by radiation. Here, transmutation elements Re and Ta atoms have been selected to assess the effects on property of vacancy and vacancy cluster in W material via first-principles calculations. The formation energy indicates mono-vacancy is more likely to form in W-Re system than pure W and W-Ta system. Both Re and Ta have reduced the diffusion barrier energy in the mono-vacancy migration. The calculation presents that vacancy cluster prefers to grow up by combining another vacancy cluster relative to a single mono-vacancy. Re is favorable to the nucleation and growth of vacancies clusters, while Ta has a suppressive effect on the aggregation of small vacancy cluster. The emphasis analysis is obtained according the volumetric dependent strain. Vacancy dissociation calculations show that the dissociation of vacancy clusters is easier to begin with a single vacancy dissociation process.

cond-mat.mtrl-sci↗

arXiv2019

Effect of tungsten on vacancy behaviors in Ta-W alloys from first-principles

Alloying elements play an important role in the design of plasma facing materials with good comprehensive properties. Based on first-principles calculations, the stability of alloying element W and its interaction with vacancy defects in Ta-W alloys are studied. The results show that W tends to distribute dispersedly in Ta lattice, and is not likely to form precipitation even with the coexistence of vacancy. The aggregation behaviors of W and vacancy can be affected by their concentration competition. The increase of W atoms has a negative effect on the vacancy clustering, as well as delays the vacancy nucleation process, which is favorable to the recovery of point defects. Our results are in consistent with the defect evolution observed in irradiation experiments in Ta-W alloys. Our calculations suggest that Ta is a potential repairing element that can be doped into Ta-based materials to improve their radiation resistance.

cond-mat.mtrl-sci↗