LEAP: LLM-powered End-to-end Automatic Library for Processing Social Science Queries on Unstructured Data
Summary: LEAP end-to-end processes vague social-science queries over unstructured data by filtering nondeterministic requests, selecting ML annotation functions, and generating executable code. On QUIET-ML, it achieves 92% pass@1 at $1.06/query. (summarized by gpt-5.6-luna on Jul 24 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Chuxuan Hu (University of Illinois Urbana-Champaign)
- 2. Austin Peters (University of Chicago)
- 3. Daniel Kang (University of Illinois Urbana-Champaign)
BibTeX Citation
@article{hu_vldb25,
title = {{LEAP: LLM-powered End-to-end Automatic Library for Processing Social Science Queries on Unstructured Data}},
author = {Hu, Chuxuan and Peters, Austin and Kang, Daniel},
journal = {PVLDB},
series = {{VLDB} '25},
volume = {18},
number = {2},
pages = {253--264},
doi = {10.14778/3705829.3705843},
url = {https://doi.org/10.14778/3705829.3705843},
year = {2025}
}Incoming Citations (Sorted by Pagerank)
Showing 4 of 4 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 6,615 | Multi-Objective Agentic Rewrites for Unstructured Data Processing | 2026 | VLDB | 5.7164252e-05 |
| 10,575 | ScaleDoc: Scaling LLM-based Predicates over Large Document Collections | 2026 | SIGMOD | 4.9605343e-05 |
| 10,634 | Drama: Unifying Data Retrieval and Analysis for Open-Domain Analytic Queries | 2026 | SIGMOD | 4.9605343e-05 |
| 11,002 | IMLane: Composable Framework for Efficient AI Function Execution in Database Engine | 2026 | VLDB | 4.9605343e-05 |
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 4 of 4 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 163 | Text-to-SQL Empowered by Large Language Models: A Benchmark Evaluation | 2024 | VLDB | 0.00027456647 |
| 1,209 | CatSQL: Towards Real World Natural Language to SQL Applications | 2023 | VLDB | 0.00011499224 |
| 2,118 | NL2SQL is a solved problem... Not! | 2024 | CIDR | 9.0008935e-05 |
| 2,479 | A Relational Approach to Incrementally Extracting and Querying Structure in Unstructured Data | 2007 | VLDB | 8.3884752e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 11,078 | iPDB: SQL with ML and LLM Predicates (Towards a Database Engine for AI) | 2026 | VLDB |
| 2 | 5,908 | Unify: A System For Unstructured Data Analytics | 2025 | VLDB |
| 3 | 8,707 | Relational Deep Dive: Error-Aware Queries Over Unstructured Data | 2026 | VLDB |
| 4 | 10,766 | NL2SQLBench: A Modular Benchmarking Framework for LLM-Enabled NL2SQL Solutions | 2026 | VLDB |
| 5 | 11,501 | LEAP: A Low-cost Spark SQL Query Optimizer using Pairwise Comparison | 2025 | VLDB |
| 6 | 8,708 | Unstructured Data Analysis Using LLMs: A Comprehensive Benchmark | 2026 | VLDB |
| 7 | 4,583 | QUEST: Query Optimization in Unstructured Document Analysis | 2025 | VLDB |
| 8 | 7,894 | ST-Raptor: LLM-Powered Semi-Structured Table Question Answering | 2026 | SIGMOD |
| 9 | 3,914 | AOP: Automated and Interactive LLM Pipeline Orchestration for Answering Complex Queries | 2025 | CIDR |
| 10 | 10,500 | Bridging the Gap: Cardinality Estimation for Semantic Queries on Unstructured Data | 2026 | SIGMOD |