Jianqiao Lu 陆剑桥
About me
Currently pursuing a PhD under the supervision of Prof. Zhiyi Huang, with a primary research focus on Large Language Models and Online Matching Theory.
Email jqlu@cs.hku.hk
Collaboration
I feel fortunate and empowered to work closely with Dr. Zhijiang Guo, Dr. Zhengying Liu, Dr. Yinya Huang, and Yingjia Wan. Our collaboration focuses on advancing topics in both formal and informal mathematical reasoning for Large Language Models, driving innovation and excellence in these fields. .
News
- Sep. 2024: (First Author) "AutoPSV: Automated Process-Supervised Verifier" , was accepted to NeurIPS 2024.
- Sep. 2024: "Proving Theorems Recursively" , was accepted to NeurIPS 2024.
- Sep. 2024: "Mr.Bean: A Comprehensive Meta-Reasoning Benchmark for Analyzing Large Language Models" , was accepted to NeurIPS 2024.
- Sep. 2024: "Online Matching Meets Sampling Without Replacement" , was accepted to WINE 2024.
- Sep. 2024: "FVEL: Interactive Formal Verification Environment with Large Language Models via Theorem Proving" , was accepted to NeurIPS 2024 (Datasets and Benchmarks Track). Project
- May 2024: "Planning, Creation, Usage: Benchmarking LLMs for Comprehensive Tool Utilization in Real-World Complex Scenarios" , was accepted to ACL 2024. Code
- Jan 2024: (First Author) "Improving End-to-End Speech Processing by Efficient Text Data Utilization with Latent Synthesis" , was accepted to EMNLP 2023.
Preprint
Large Language Models
- Oct 2023: (First Author) "SELF: Self-Evolution with Language Feedback", was released on arXiv. (A self-evolution framework for autonomous self-improvement of large language models through self-feedback and iterative self-refinement.)
- Jan 2024: (First Author) "YODA: Teacher-Student Progressive Learning for Language Models", was released on arXiv. (A learning framework that mimics human progressive learning to enhance language model fine-tuning.)
- Jan 2024: (Equal Contribution) "MHPP: Exploring the Capabilities and Limitations of Language Models Beyond Basic Code Generation", was released on arXiv. (A benchmark to rigorously assess LLMs' code generation abilities, revealing significant gaps in existing benchmarks and providing new insights into LLMs' strengths and weaknesses.)
- May 2024: (First Author) "Process-Driven Autoformalization in Lean 4" , was released on arXiv.
- May 2024: "DialogBench: An Interactive Benchmark for Humanized Dialogue Agent" , will be soon released on arXiv.
- June 2024: (First Author) "FormalAlign: Automated Alignment Evaluation in Autoformalization" was released on arXiv.
Office
LG101, Chow Yei Ching Building, Department of Computer Science, The University of Hong Kong.