Language models are robotic planners: reframing plans as goal refinement graphs (2407.15677v1)
Abstract: Successful application of LLMs to robotic planning and execution may pave the way to automate numerous real-world tasks. Promising recent research has been conducted showing that the knowledge contained in LLMs can be utilized in making goal-driven decisions that are enactable in interactive, embodied environments. Nonetheless, there is a considerable drop in correctness of programs generated by LLMs. We apply goal modeling techniques from software engineering to LLMs generating robotic plans. Specifically, the LLM is prompted to generate a step refinement graph for a task. The executability and correctness of the program converted from this refinement graph is then evaluated. The approach results in programs that are more correct as judged by humans in comparison to previous work.
- Pddl— the planning domain definition language. Technical Report, Tech. Rep., 1998.
- Do as i can, not as i say: Grounding language in robotic affordances. In 6th Annual Conference on Robot Learning, 2022.
- Language models are few-shot learners. Advances in neural information processing systems, 33:1877–1901, 2020.
- Strips: A new approach to the application of theorem proving to problem solving. Artificial intelligence, 2(3-4):189–208, 1971.
- GitHub. GitHub Copilot - November 30th Update. https://github.blog/changelog/2023-11-30-github-copilot-november-30th-update/, 2023a. Accessed: 2024-06-08.
- GitHub. Using GitHub Copilot code suggestions in your editor. https://docs.github.com/en/copilot/using-github-copilot/using-github-copilot-code-suggestions-in-your-editor, 2024b. Accessed: 2024-06-08.
- Language models as zero-shot planners: Extracting actionable knowledge for embodied agents. In International Conference on Machine Learning, pages 9118–9147. PMLR, 2022.
- James Herring. Github copilot: The power to boost your programming productivity. https://techcommunity.microsoft.com/t5/microsoft-learn/github-copilot-the-power-to-boost-your-programming-productivity/m-p/3858851, 2023. Accessed: 2024-06-08.
- Axel van Lamsweerde. Requirements Engineering: From System Goals to UML Models to Software Specifications. W. Ross MacDonald School Resource Services Library, 2018.
- Code as policies: Language model programs for embodied control. In 2023 IEEE International Conference on Robotics and Automation (ICRA), pages 9493–9500. IEEE, 2023.
- Language models of code are few-shot commonsense learners. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 1384–1403, 2022.
- OpenAI. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023.
- OpenAI. API Reference - OpenAI API. https://platform.openai.com/docs/api-reference/chat, 2024. Accessed:2024-06-08.
- Virtualhome: Simulating household activities via programs. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 8494–8502, 2018.
- Toolformer: Language models can teach themselves to use tools. Advances in Neural Information Processing Systems, 36, 2024.
- Progprompt: Generating situated robot task plans using large language models. In 2023 IEEE International Conference on Robotics and Automation (ICRA), pages 11523–11530. IEEE, 2023.
- Chatgpt for robotics: Design principles and model abilities. IEEE Access, 2024.
- Neural task programming: Learning to generalize across hierarchical tasks. In 2018 IEEE international conference on robotics and automation (ICRA), pages 3795–3802. IEEE, 2018.
- Regression planning networks. Advances in neural information processing systems, 32, 2019.
- From text to motion: Grounding gpt-4 in a humanoid robot” alter3”. arXiv preprint arXiv:2312.06571, 2023.
- Ateeq Sharfuddin (3 papers)
- Travis Breaux (4 papers)