Make-An-Agent：一种具有行为引发扩散的通用策略网络生成器。

Make-An-Agent: A Generalizable Policy Network Generator with Behavior-Prompted Diffusion

July 15, 2024

作者: Yongyuan Liang, Tingqiang Xu, Kaizhe Hu, Guangqi Jiang, Furong Huang, Huazhe Xu

cs.AI

摘要

我们能否仅凭所需行为的一次演示作为提示，像从文本描述中创建图像一样轻松地为代理生成控制策略？在本文中，我们提出了Make-An-Agent，这是一种新颖的策略参数生成器，利用条件扩散模型的强大功能进行行为到策略的生成。在行为嵌入的指导下，编码轨迹信息，我们的策略生成器合成潜在参数表示，然后可以解码为策略网络。在策略网络检查点及其对应轨迹上进行训练后，我们的生成模型展示了在多个任务上的出色多功能性和可扩展性，并且在未见任务上具有强大的泛化能力，能够仅凭少量演示输入输出表现良好的策略。我们展示了它在各种领域和任务上的有效性和效率，包括不同目标、行为，甚至跨不同机器人操作器的情况。除了模拟外，我们还将Make-An-Agent生成的策略直接部署到真实世界的机器人上进行运动任务。

English

Can we generate a control policy for an agent using just one demonstration of desired behaviors as a prompt, as effortlessly as creating an image from a textual description? In this paper, we present Make-An-Agent, a novel policy parameter generator that leverages the power of conditional diffusion models for behavior-to-policy generation. Guided by behavior embeddings that encode trajectory information, our policy generator synthesizes latent parameter representations, which can then be decoded into policy networks. Trained on policy network checkpoints and their corresponding trajectories, our generation model demonstrates remarkable versatility and scalability on multiple tasks and has a strong generalization ability on unseen tasks to output well-performed policies with only few-shot demonstrations as inputs. We showcase its efficacy and efficiency on various domains and tasks, including varying objectives, behaviors, and even across different robot manipulators. Beyond simulation, we directly deploy policies generated by Make-An-Agent onto real-world robots on locomotion tasks.