Meihao Fan
Ph.D. student at Renmin University of China 
Posts by Collection
people
projects
BATCHER: Cost-Effective LLM Inference for Entity Resolution
Published:
A cost-effective in-context learning framework for entity resolution.
AutoPrep: Multi-Agent Data Preparation Framework
Published:
A multi-agent data preparation framework with planner and programmer agents.
DeepPrep: Verifiable Task Synthesis for Agent Training
Published:
Project Lead; First Author · Executable supervision, reversible corruption, multi-turn RL. VLDB 2026.
CoDA-Bench & TACO: Scalable, Verifiable Task Construction
Co-author · Task synthesis, execution verification, difficulty refinement. ICML / VLDB 2026.
DeepAnalyze: Scalable Training-Trajectory Synthesis
Project Lead · Synthetic data, quality filtering, curriculum learning, agentic RL. ICML 2026.
EvoCost & SkillAdam: Agent Self-Evolution
EvoCost: Lead, First Author · SkillAdam: Co-author. ICLR 2027 submissions.
Interactive Memory & Recursive Self-Improvement (RSI) for World Models
Project Lead · World model evaluation, scalable data curation, training, self-improvement. Ongoing.
publications
Few‐Shot Multihop Question Answering over Knowledge Base
Published in Wireless Communications and Mobile Computing, 2021
This paper is about Question Answering over Knowledge Base.
Recommended citation: Fan, Meihao, Lei Zhang, Siyao Xiao, and Yuru Liang. "Few-shot multi-hop question answering over knowledge base." arXiv preprint arXiv:2112.11909 (2021). https://onlinelibrary.wiley.com/doi/abs/10.1155/2022/8045535
Cost-Effective In-Context Learning for Entity Resolution: A Design Space Exploration
Published in ICDE 2024, 2024
This paper is about LLM for Entity Resolution.
Recommended citation: Meihao Fan, Xiaoyue Han, Ju Fan, Chengliang Chai, Nan Tang, Guoliang Li, Xiaoyong Du. "Cost-Effective In-Context Learning for Entity Resolution: A Design Space Exploration" ICDE 2024. https://fmh1art.github.io/files/BatchER-ICDE2024.pdf
AutoPrep: Natural Language Question-Aware Data Preparation with a Multi-Agent Framework
Published in VLDB 2025, 2025
This paper proposes AutoPrep, a multi-agent framework for data preparation.
Recommended citation: Meihao Fan, Ju Fan, Nan Tang, Lei Cao, Guoliang Li, Xiaoyong Du. "AutoPrep: Natural Language Question-Aware Data Preparation with a Multi-Agent Framework" VLDB 2025. https://www.vldb.org/pvldb/vol18/p3504-fan.pdf
Reward-SQL: Boosting Text-to-SQL via Stepwise Execution-Aware Reasoning and Process-Supervised Rewards
Published in SIGMOD 2026, 2026
This paper proposes Reward-SQL, a framework for improving Text-to-SQL reasoning using process-supervised reward models.
Recommended citation: Yuxin Zhang, Meihao Fan, Ju Fan, Mingyang Yi, Yuyu Luo, Jian Tan, Guoliang Li. "Reward-SQL: Boosting Text-to-SQL via Stepwise Execution-Aware Reasoning and Process-Supervised Rewards" SIGMOD 2026. https://arxiv.org/abs/2505.04671
DeepPrep: An LLM-Powered Agentic System for Autonomous Data Preparation
Published in VLDB 2026, 2026
This paper proposes DeepPrep, an LLM-powered agentic system for autonomous data preparation.
Recommended citation: Meihao Fan, Ju Fan, Yuxin Zhang, Shaolei Zhang, Xiaoyong Du, Jie Song, Peng Li, Fuxin Jiang, Tieying Zhang, Jianjun Chen. "DeepPrep: An LLM-Powered Agentic System for Autonomous Data Preparation" VLDB 2026. https://www.vldb.org/pvldb/vol19/p3371-fan.pdf
TACO: A Benchmark for Open-Domain Text-to-SQL with Ambiguous and Cross-Database Queries
Published in VLDB 2026, 2026
This paper proposes TACO, a benchmark for Open-Domain Text-to-SQL.
Recommended citation: Chao Deng, Ju Fan, Yuyu Luo, Qinliang Xue, Meihao Fan, Yuxin Zhang, Min Zhang, Xiaofeng Jia, Jing Zhang, Xiaoyong Du. "TACO: A Benchmark for Open-Domain Text-to-SQL with Ambiguous and Cross-Database Queries" VLDB 2026. https://arxiv.org/abs/2606.14201
DeepAnalyze: Agentic Large Language Models for Autonomous Data Science
Published in ICML 2026, 2026
This paper explores agentic large language models for autonomous data science.
Recommended citation: Shaolei Zhang, Ju Fan, Meihao Fan, Guoliang Li, Xiaoyong Du. "DeepAnalyze: Agentic Large Language Models for Autonomous Data Science" ICML 2026. https://arxiv.org/abs/2510.16872
CoDA-Bench: Can Code Agents Handle Data-Intensive Tasks?
Published in ICML 2026, 2026
This paper introduces CoDA-Bench to evaluate code agents on data-intensive tasks.
Recommended citation: Yuxin Zhang, Ju Fan, Meihao Fan, Shaolei Zhang, Xiaoyong Du. "CoDA-Bench: Can Code Agents Handle Data-Intensive Tasks?" ICML 2026. https://arxiv.org/abs/2606.15300
software
Predictor identifier: Nonparametric PREDiction (NPRED)
The open-source R package NPRED is used to identify the meaningful predictors to the response from a large set of potential predictors.
WASP: Wavelet System Prediction
The open-source software WASP is used for system modeling and prediction.
WQM: Wavelet-based Quantile Mapping
The open-source software WQM is used for post-processing numerical weather prediction.
synthesis: Generate Synthetic Data from Statistical Models
Generate synthetic time series from commonly used statistical models, including linear, nonlinear and chaotic systems.