NOTES

  • This paper is to appear at ICRA 2025!
  • This paper builds upon our prior work on FOON-bootstrapped task and motion planning by naturally interfacing hierarchical planning with large language models (LLMs).
    • Language models are all the rage nowadays and are excellent at doing few-shot learning and deployment.
    • For this reason, we use them to construct object-level plans (in the form of a FOON), thus overcoming the limitation we have encountered before in hand-crafting FOONs from video.
  • We demonstrate how leveraging automated solving with language models enables better performance on table-top block manipulation tasks.

Citation

D. Paulius, A. Agostini, B. Quartey, G. Konidaris (2025). " Bootstrapping Object-level Planning with Large Language Models". In: 2025 IEEE International Conference on Robotics and Automation (ICRA). IEEE.