1. LLMs as Translators, Not Thinkers: Structured Output Enables Stronger NP-Hard Problem Solving
2. AUTOCT: Automating Interpretable Clinical Trial Prediction with LLM Agents
Peer-reviewed papers:
1. Devil's Advocate: Anticipatory Reflection for LLM Agents
2. Event Causality Identification with Synthetic Control
3. BLINK: Multimodal Large Language Models Can See but Not Perceive
4. Event Semantic Classification in Context
5. Are All Steps Equally Important? Benchmarking Essentiality Detection of Events
6. Generic Temporal Reasoning with Differential Analysis and Explanation
7. Extracting or Guessing? Improving Faithfulness of Event Temporal Relation Extraction
8. Zero-Shot On-the-Fly Event Schema Induction
9. Capturing the Content of a Document through Complex Event Identification
Research Experience
1. Research Assistant at UPenn (Sep 2019 - Present) - Event-Centric NLP/NLU; LLM Reasoning.
2. Applied Scientist Intern at Amazon AWS (May 2024 - Aug 2024) - Ambiguity Driven Exploration for Agent Planning.
3. Student Researcher at Google DeepMind (Feb 2024 - May 2024) - Anticipatory Reflection for LLM Agents.
4. Research Intern at ByteDance (Jun 2023 - Aug 2023) - Building E-Commerce Intelligent Assistant with LLMs.
5. Research Intern at AI2 (Sep 2022 - Apr 2023) - Event Causality Identification with Synthetic Control and LM Inversion.
6. Research Intern at Tencent AI Lab (May 2022 - Aug 2022) - Event Semantic Classification in Context.
7. Research Assistant at SJTU (Dec 2017 - May 2019) - Pose Estimation and Tracking.
Education
PhD student at the Cognitive Computation Group, UPenn, under Prof. Dan Roth; Bachelor's degree from Shanghai Jiao Tong University, supervised by Prof. Cewu Lu.
Background
Research interests: LLM reasoning and planning, Event-Centric NLP/NLU. Undergraduate research at Shanghai Jiao Tong University on pose estimation and tracking in Computer Vision.
Miscellany
Interned at Google DeepMind, Amazon AWS, ByteDance, AI2, Tencent AI Lab, and Goldman Sachs.