From Multimodal LLMs to Generalist Embodied Agents: Methods and Lessons
This work addresses the challenge of creating unified AI agents for diverse real-world applications, representing an incremental advance by building on existing MLLMs with new training methods.
The paper tackles the problem of adapting Multimodal Large Language Models (MLLMs) to function as Generalist Embodied Agents (GEAs) across domains like Embodied AI and Games, achieving strong generalization performance on unseen tasks compared to other models.
We examine the capability of Multimodal Large Language Models (MLLMs) to tackle diverse domains that extend beyond the traditional language and vision tasks these models are typically trained on. Specifically, our focus lies in areas such as Embodied AI, Games, UI Control, and Planning. To this end, we introduce a process of adapting an MLLM to a Generalist Embodied Agent (GEA). GEA is a single unified model capable of grounding itself across these varied domains through a multi-embodiment action tokenizer. GEA is trained with supervised learning on a large dataset of embodied experiences and with online RL in interactive simulators. We explore the data and algorithmic choices necessary to develop such a model. Our findings reveal the importance of training with cross-domain data and online RL for building generalist agents. The final GEA model achieves strong generalization performance to unseen tasks across diverse benchmarks compared to other generalist models and benchmark-specific approaches.