I am a second-year PhD student at the Hong Kong University of Science and Technology (Guangzhou), honored to be supervised by Professor Yuyu Luo. Currently, I am working as a research intern in the Database Group led by Professor Guoliang Li at Tsinghua University. Previously, I obtained my bachelor’s degree from the Southern University of Science and Technology. My current research interests focus on Text-to-SQL and Data Agent, which are intelligent systems that can autonomously handle data-related tasks such as query generation, data analysis, and visualization through natural language interaction.

News

2025.10 We proposed A Survey of Data Agents: Emerging Paradigm or Overstated Hype?, introducing the first systematic hierarchical taxonomy for data agents with six levels (L0-L5) that delineate progressive shifts in autonomy, from manual operations to fully autonomous data agents. This survey clarifies capability boundaries and responsibility allocation, offering a structured review of existing research and a forward-looking roadmap. Ranked Top-3 in Hugging Face Daily Papers!
2025.10 We proposed DeeyEye-SQL, a software-engineering-inspired Text-to-SQL framework, achieving 73.5% EX on BIRD-Dev and 89.8% EX on Spider-Test using a ~30B LLM without any fine-tuning!
2025.09 Our nvBench 2.0 paper has been accepted by NIPS'25, a new benchmark designed to evaluate Text2VIS systems in scenarios involving ambiguous queries.
2025.07 Our DeepVIS paper has been accepted by VIS'25, an interactive visual interface that tightly integrates with the CoT reasoning process, allowing users to inspect reasoning steps, identify errors, and make targeted adjustments to improve visualization outcomes.
2025.07 Our NL2SQL-Survey paper has been accepted by TKDE'25! For a comprehensive overview of the latest Text-to-SQL techniques and practical guidance, we warmly invite you to read our continuously updated NL2SQL Handbook.
2025.07 Our EllieSQL paper has been accepted by COLM'25, a complexity-aware routing framework that assigns queries to suitable SQL generation pipelines based on estimated complexity.
2025.05 Our NL2SQL-BUGs paper has been accepted by KDD'25, the first benchmark specifically designed to detect and categorize semantic errors in NL2SQL translation.
2025.05 Our Alpha-SQL paper has been accepted by ICML'25, see you in Vancouver, Canada!
2025.04 We proposed Advances and Challenges in Foundation Agents, a survey covers the design, evaluation, and improvement of intelligent agents based on modular, brain-inspired architectures, focusing on self-enhancement, multi-agent collaboration, and safety in AI systems.
2025.04 We proposed EllieSQL, a complexity-aware routing framework that assigns queries to suitable SQL generation methods based on estimated complexity.
2025.03 We proposed NL2SQL-BUGs, a new benchmark dedicated to detecting and categorizing semantic errors in NL2SQL translation.
2025.03 We proposed nvBench 2.0, a new benchmark designed to evaluate NL2VIS systems in scenarios involving ambiguous queries.
2025.01 We proposed Alpha-SQL, the o1 moment for NL2SQL!
2025.01 Paper Augmenting Realistic Charts with Virtual Overlays has been accepted by CHI'25.
2025.01 I was awarded the Merit Prize for the 2024 DSA Excellent Research Award!
2024.09 Paper Are Large Language Models Good Statisticians? has been accepted by NIPS'24.
2024.06 Paper The Dawn of Natural Language to SQL: Are We Fully Ready? has been accepted by VLDB'24.
2024.04 Paper Efficient Deep Spiking Multilayer Perceptrons With Multiplication-Free Inference has been accepted by TNNLS'24.