Demonstrating TOFFEE: A Learned System for Synthesizing Data Agent Trajectories at Scale
Summary: TOFFEE synthesizes scalable, high-quality data-agent trajectories for heterogeneous environments using MCTS, adaptive model selection, and cross-task prefix reuse. Supports both domain-specific SFT and demonstration-augmented ICL for complex analytical workflows. (summarized by gpt-5.6-luna on Aug 28 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Ziting Wang (Nanyang Technological University)
- 2. Yin Li (Nanyang Technological University)
- 3. Zuhao Yang (Nanyang Technological University)
- 4. Xiuchang Li (Huawei)
- 5. Jiale Bai (Industrial and Commercial Bank of China Limited)
- 6. Gao Cong (Nanyang Technological University)
BibTeX Citation
@article{wang_vldb26,
title = {{Demonstrating TOFFEE: A Learned System for Synthesizing Data Agent Trajectories at Scale}},
author = {Wang, Ziting and Li, Yin and Yang, Zuhao and Li, Xiuchang and Bai, Jiale and Cong, Gao},
journal = {PVLDB},
series = {{VLDB} '26},
volume = {19},
number = {12},
pages = {4738--4741},
doi = {10.14778/3827998.3828110},
url = {https://doi.org/10.14778/3827998.3828110},
year = {2026}
}Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 2 of 2 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 1,777 | OmniSQL: Synthesizing High-quality Text-to-SQL Data at Scale | 2025 | VLDB | 9.664552e-05 |
| 7,090 | LEAD: Iterative Data Selection for Efficient LLM Instruction Tuning | 2026 | VLDB | 5.601767e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 5,687 | DTT: An Example-Driven Tabular Transformer for Joinability by Leveraging Large Language Models | 2024 | SIGMOD |
| 2 | 3,423 | AutoTQA: Towards Autonomous Tabular Question Answering through Multi-Agent Large Language Models | 2024 | VLDB |
| 3 | 8,230 | Adda: Towards Efficient in-Database Feature Generation via LLM-based Agents | 2025 | SIGMOD |
| 4 | 13,601 | DA-Studio: An Agentic System for End-to-End Data Analysis | 2026 | VLDB |
| 5 | 11,164 | Sentence to Model: Cost-Effective Data Collection LLM Agent | 2025 | SIGMOD |
| 6 | 13,609 | Data Management for Agentic Memory: Foundations, Systems, and Challenges | 2026 | VLDB |
| 7 | 10,869 | DeepPrep: An LLM-Powered Agentic System for Autonomous Data Preparation | 2026 | VLDB |
| 8 | 5,149 | LLM for Data Management | 2024 | VLDB |
| 9 | 10,997 | Demonstration of BobFlow: Human-Agent Collaboration Using Dataflows on Data Science Tasks | 2026 | VLDB |
| 10 | 9,444 | Intelligent Agents for Data Exploration | 2024 | VLDB |