Optimizing Agent Behavior in the MiniGrid Environment Using Reinforcement Learning Based on Large Language Models
Source abstract
Abstract
Reinforcement learning is one of the most prominent research areas in the field of artificial intelligence, playing a crucial role in developing agents that autonomously make decisions in complex environments. This study proposes a method to optimize agent behavior in the MiniGrid-Empty-5x5-v0 environment using large language models (LLMs). By leveraging the natural language processing capabilities of LLMs to interpret environmental states and select appropriate actions, this research explores an approach that differs from traditional reinforcement learning methods. Experimental results confirm that LLM-based agents can effectively achieve their goals, and it is anticipated that maximizing the synergy between LLMs and reinforcement learning will contribute to the development of more intelligent and adaptable AI systems.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Reinforcement Learning in Robotics · Smart Grid Energy Management
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
School of Computer Science and Engineering, Korea University of Technology and Education, Cheonan 31253, Republic of Korea
Funding
National Research Foundation of Korea
References (17)
Show source-provided references
1. Survey on Practical Reinforcement Learning: From Imitation Learning to Offline Reinforcement Learning
2. Analysis of Changes in AI Perception Using Text Mining from AlphaGo to Chat GPT by Focusing on Legacy Media and Social Media10.22417/DPSS.2023.10.16.2.205 ↗
3. Autonomous Navigation Command Generation System for Underwater Robots Using LLM10.5302/J.ICROS.2024.24.0142 ↗
4. Zhao, W.X., and Zhou, K. (2023). A survey of large language models. arXiv.
5. Koubaa, A., Boulila, W., Ghouti, L., Alzahem, A., and Latif, S. (2023). Exploring ChatGPT capabilities and limitations: A critical review of the NLP game changer. Preprints.10.20944/preprints202303.0438.v1 ↗
6. Open AI (2023). GPT-4 Technical Report. arXiv.
7. A Study on the Efficient Response to Architectural Civil Complaints Using Large Language Models (LLM)
8. (2024, August 19). DATAVERSITY. Available online: https://www.dataversity.net/a-brief-history-of-large-language-models/.
9. Enhancing RAG-LLM Response Performance Using Ensemble Diversity
10. Bahrini, A., Khamoshifar, M., Abbasimehr, H., Riggs, R., Esmaeili, M., Aajdabadkohne, R., and Pasehvar, M. (2023). ChatGPT: Applications, opportunities, and threats. arXiv.10.1109/SIEDS58326.2023.10137850 ↗
11. Haarnoja, T., Zhou, T., Abbeel, A., and Levine, S. (2018). Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor. arXiv.
12. Deep reinforcement learning for autonomous driving: A survey10.1109/TITS.2021.3054625 ↗
13. Laskin, M., Srinivas, A., and Abbeel, P. (2020). CURL: Contrastive unsupervised representations for reinforcement learning. arXiv.
14. Bommasani, R., Hudson, D., and Adeli, A. (2021). On the opportunities and risks of foundation models. arXiv.
15. Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity
16. Wei, J., Bosma, M., and Zhao, V. (2022). Chain of thought prompting elicits reasoning in large language models. arXiv.
17. Bubeck, S., Chandrasekaran, V., and Eldan, R. (2023). Sparks of artificial general intelligence: Early experiments with GPT-4. arXiv.