Institution profile

Yale University

Academic institutionnorthamerica · us
Official website
Research library1,377linked papers
Opportunities0open roles
Selected work

Representative Papers

An Empirical Study of Automated Vulnerability Localization with Large Language Models

Mar 30, 2024arXiv.org

This work systematically evaluates the effectiveness of large language models (LLMs) for line-level vulnerability localization (AVL)—a task lacking comprehensive empirical investigation. Experiments are conducted on BigVul (C/C++) and smart contract vulnerability datasets, covering over ten code-understanding LLMs (60M–16B parameters) spanning encoder-only, encoder-decoder, and decoder-only architectures, under zero-shot, one-shot, discriminative fine-tuning, and generative fine-tuning paradigms. Key contributions include: (1) the first empirical demonstration that discriminative fine-tuning substantially outperforms existing approaches; (2) the proposal of sliding-window context partitioning and right-forward embedding to mitigate context-length limitations; and (3) strong cross-CWE and cross-project generalization, yielding significant improvements in localization accuracy and surpassing state-of-the-art methods.

25 citations2 influentialRead paper

Trademark Search, Artificial Intelligence and the Role of the Private Sector

Jan 22, 2026

This study addresses a critical gap in trademark scholarship, which has predominantly focused on consumer search costs while overlooking the substantial burdens faced by applicants and the transformative role of artificial intelligence (AI) in trademark creation and retrieval. Through empirical experiments, the paper evaluates leading AI-powered trademark search engines, assessing their performance in judging trademark similarity. It proposes a reconceptualization of the trademark analysis framework from a supply-side perspective, foregrounding AI as a central actor in trademark selection. The findings demonstrate that AI tools significantly enhance both the efficiency and accuracy of trademark searches, thereby advancing trademark law toward a more balanced paradigm that simultaneously promotes innovation incentives and institutional efficiency.

9 citationsRead paper

Agentic Reasoning for Large Language Models

Jan 18, 2026

This work addresses the limited capacity of large language models (LLMs) to plan, act, and learn through sustained interaction in open, dynamic environments. To overcome this, the authors propose a three-tiered reasoning framework that treats LLMs as autonomous agents, unifying single-agent foundational reasoning, self-evolution, and multi-agent collaboration within a coherent paradigm. The framework orchestrates structured interactions, incorporates memory mechanisms, enables tool use, and integrates both reinforcement learning and supervised fine-tuning. It explicitly distinguishes between in-context reasoning and post-training optimization pathways, systematically coupling cognition with action. Empirical validation across diverse domains—including scientific discovery, robotics, healthcare, autonomous research, and mathematics—demonstrates its effectiveness and highlights promising future directions such as personalized interaction, long-horizon engagement, world modeling, and scalable multi-agent training.

7 citations1 influentialRead paper

Scheduling With Time Discounts

Feb 13, 2024

This paper studies the online weighted task scheduling problem with time-decaying values in financial settings, where future rewards are discounted at rate δ ∈ [0,1) and arrivals are uncertain; the objective is to maximize the total present value. We propose the first deterministic memoryless algorithm that is optimal for δ ≤ 0.77 and prove its competitive ratio matches the theoretical upper bound for this class of algorithms. Furthermore, we design a randomized algorithm that strictly surpasses the deterministic competitive ratio upper bound—thereby completing the precise characterization of the competitive ratio for discounted scheduling. Our theoretical results are directly applicable to real-world financial systems, such as blockchain transaction scheduling, and provide a unified modeling framework and provably optimal algorithms for online resource allocation under discounted utility.

6 citationsRead paper

Using Advanced LLMs to Enhance Smaller LLMs: An Interpretable Knowledge Distillation Approach

Aug 13, 2024arXiv.org

In edge- and on-premises-deployed, goal-oriented customer service dialogue scenarios, existing large language model (LLM) solutions struggle to balance performance, controllability, and cost—proprietary models (e.g., GPT-4) incur high licensing fees and lack self-hosting capability, while open-source lightweight models suffer from insufficient capability. Method: We propose “policy distillation,” a novel black-box, interpretable knowledge transfer paradigm comprising two stages: scenario-aware generation and policy optimization. It constructs an auditable, transferable prompt policy library—eliminating reliance on parameter fine-tuning or response imitation—by integrating black-box API invocation, scenario-driven policy induction, and automated prompt engineering. Contribution/Results: Experiments demonstrate substantial improvement in user satisfaction for lightweight LLMs on customer service tasks; the distilled policies exhibit strong generalization across models and tasks; and built-in human review support enhances safety and operational controllability.

4 citationsRead paper
Recent publications

Latest Papers