WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environment
This work addresses efficiency challenges in reinforcement learning for AI agents, though it is incremental as it builds on existing model-based and code-writing approaches.
The authors tackled the problem of sample and compute inefficiency in reinforcement learning by introducing WorldCoder, a model-based LLM agent that builds a Python program as a world model through environment interactions, achieving improved sample efficiency over deep RL and compute efficiency over ReAct-style agents in gridworlds and task planning.
We give a model-based agent that builds a Python program representing its knowledge of the world based on its interactions with the environment. The world model tries to explain its interactions, while also being optimistic about what reward it can achieve. We define this optimism as a logical constraint between a program and a planner. We study our agent on gridworlds, and on task planning, finding our approach is more sample-efficient compared to deep RL, more compute-efficient compared to ReAct-style agents, and that it can transfer its knowledge across environments by editing its code.