ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2506.11045
13
0

Procedural Environment Generation for Tool-Use Agents

21 May 2025
Michael Sullivan
Mareike Hartmann
Alexander Koller
    SyDa
ArXiv (abs)PDFHTML
Main:8 Pages
3 Figures
Bibliography:3 Pages
6 Tables
Appendix:5 Pages
Abstract

Although the power of LLM tool-use agents has ignited a flurry of recent research in this area, the curation of tool-use training data remains an open problem−-−especially for online RL training. Existing approaches to synthetic tool-use data generation tend to be non-interactive, and/or non-compositional. We introduce RandomWorld, a pipeline for the procedural generation of interactive tools and compositional tool-use data. We show that models tuned via SFT and RL on synthetic RandomWorld data improve on a range of tool-use benchmarks, and set the new SoTA for two metrics on the NESTFUL dataset. Further experiments show that downstream performance scales with the amount of RandomWorld-generated training data, opening up the possibility of further improvement through the use of entirely synthetic data.

View on arXiv
@article{sullivan2025_2506.11045,
  title={ Procedural Environment Generation for Tool-Use Agents },
  author={ Michael Sullivan and Mareike Hartmann and Alexander Koller },
  journal={arXiv preprint arXiv:2506.11045},
  year={ 2025 }
}
Comments on this paper