2000 character limit reached
Bootstrapping Cognitive Agents with a Large Language Model
Published 25 Feb 2024 in cs.AI and cs.CL | (2403.00810v1)
Abstract: LLMs contain noisy general knowledge of the world, yet are hard to train or fine-tune. On the other hand cognitive architectures have excellent interpretability and are flexible to update but require a lot of manual work to instantiate. In this work, we combine the best of both worlds: bootstrapping a cognitive-based model with the noisy knowledge encoded in LLMs. Through an embodied agent doing kitchen tasks, we show that our proposed framework yields better efficiency compared to an agent based entirely on LLMs. Our experiments indicate that LLMs are a good source of information for cognitive architectures, and the cognitive architecture in turn can verify and update the knowledge of LLMs to a specific domain.
- Do as i can, not as i say: Grounding language in robotic affordances. arXiv preprint arXiv:2204.01691.
- Anderson, J. R. 2009. How can the human mind occur in the physical universe? Oxford University Press.
- On evaluation of embodied navigation agents. arXiv preprint arXiv:1807.06757.
- Habitat Rearrangement Challenge 2022. https://aihabitat.org/challenge/rearrange˙2022.
- Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback. arXiv preprint arXiv:2307.15217.
- Towards a unified agent with foundation models. In Workshop on Reincarnating Reinforcement Learning at ICLR 2023.
- Palm-e: An embodied multimodal language model. arXiv preprint arXiv:2303.03378.
- Dreamcoder: Bootstrapping inductive program synthesis with wake-sleep library learning. In Proceedings of the 42nd acm sigplan international conference on programming language design and implementation, 835–850.
- Inferring a Cognitive Architecture from Multitask Neuroimaging Data: A Data-Driven Test of the Common Model of Cognition Using Granger Causality. Topics in Cognitive Science, 14(4): 845–859.
- Language models as zero-shot planners: Extracting actionable knowledge for embodied agents. In International Conference on Machine Learning, 9118–9147. PMLR.
- Learning Hierarchical Symbolic Representations to Support Interactive Task Learning and Knowledge Transfer. In IJCAI, 6095–6102.
- Integrating Diverse Knowledge Sources for Online One-shot Learning of Novel Tasks. arXiv:2208.09554.
- AI2-THOR: An Interactive 3D Environment for Visual AI. arXiv.
- Laird, J. E. 2017. SOAR 9.6.0 Tutorial.
- Laird, J. E. 2022. Introduction to Soar. arXiv preprint arXiv:2205.03854.
- Interactive task learning. IEEE Intelligent Systems, 32(4): 6–21.
- A standard model of the mind: Toward a common computational framework across artificial intelligence, cognitive science, neuroscience, and robotics. Ai Magazine, 38(4): 13–26.
- Code as policies: Language model programs for embodied control. In 2023 IEEE International Conference on Robotics and Automation (ICRA), 9493–9500. IEEE.
- Improving Knowledge Extraction from LLMs for Robotic Task Learning through Agent Analysis. arXiv preprint arXiv:2306.06770.
- LLM+ P: Empowering Large Language Models with Optimal Planning Proficiency. arXiv preprint arXiv:2304.11477.
- Lost in the middle: How language models use long contexts. arXiv preprint arXiv:2307.03172.
- Language models of code are few-shot commonsense learners. arXiv preprint arXiv:2210.07128.
- A survey of explainable reinforcement learning. arXiv preprint arXiv:2202.08434.
- A Demonstration of Compositional, Hierarchical Interactive Task Learning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, 13203–13205.
- Newell, A. 1994. Unified theories of cognition. Harvard University Press.
- Demystifying GPT Self-Repair for Code Generation. arXiv:2306.09896.
- OpenAI. 2023. GPT-4 Technical Report. arXiv:2303.08774.
- Generative agents: Interactive simulacra of human behavior. arXiv preprint arXiv:2304.03442.
- Embodied Executable Policy Learning with Language-based Scene Summarization. arXiv preprint arXiv:2306.05696.
- Tidee: Tidying up novel rooms using visuo-semantic commonsense priors. In European Conference on Computer Vision, 480–496. Springer.
- ProgPrompt: Generating Situated Robot Task Plans using Large Language Models. In International Conference on Robotics and Automation (ICRA).
- Llm-planner: Few-shot grounded planning for embodied agents with large language models. arXiv preprint arXiv:2212.04088.
- Reinforcement learning: An introduction. MIT press.
- Towards Modeling and Influencing the Dynamics of Human Learning. In Proceedings of the 2023 ACM/IEEE International Conference on Human-Robot Interaction, 350–358.
- A Simple Approach for Visual Room Rearrangement: 3D Mapping and Semantic Search. In The Eleventh International Conference on Learning Representations.
- ChatGPT for Robotics: Design Principles and Model Abilities. Technical Report MSR-TR-2023-8, Microsoft.
- Voyager: An Open-Ended Embodied Agent with Large Language Models. arXiv:2305.16291.
- Chain-of-thought prompting elicits reasoning in large language models. Advances in Neural Information Processing Systems, 35: 24824–24837.
- Tidybot: Personalized robot assistance with large language models. arXiv preprint arXiv:2305.05658.
- Language Models Meet World Models: Embodied Experiences Enhance Language Models. arXiv preprint arXiv:2305.10626.
- Translating natural language to planning goals with large-language models. arXiv preprint arXiv:2302.05128.
- AmadeusGPT: a natural language interface for interactive animal behavioral analysis. arXiv preprint arXiv:2307.04858.
- Large Language Models as Zero-Shot Human Models for Human-Robot Interaction. arXiv:2303.03548.
- Ghost in the Minecraft: Generally Capable Agents for Open-World Enviroments via Large Language Models with Text-based Knowledge and Memory. arXiv preprint arXiv:2305.17144.
- Universal and Transferable Adversarial Attacks on Aligned Language Models. arXiv:2307.15043.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.
