“Imitating Human Behavior With Diffusion Models”, Tim Pearce, Tabish Rashid, Anssi Kanervisto, Dave Bignell, Mingfei Sun, Raluca Georgescu, Sergio Valcarcel Macua, Shan Zheng Tan, Ida Momennejad, Katja Hofmann, Sam Devlin2023-01-25 (diffusion model, imitation learning)⁠:

[cf. Wang et al 2022] Diffusion models have emerged as powerful generative models in the text-to-image domain.

This paper studies their application as observation-to-action models for imitating human behavior in sequential environments. Human behavior is stochastic and multimodal, with structured correlations between action dimensions. Meanwhile, standard modeling choices in behavior cloning are limited in their expressiveness and may introduce bias into the cloned policy. We begin by pointing out the limitations of these choices. We then propose that diffusion models are an excellent fit for imitating human behavior, since they learn an expressive distribution over the joint action space.

We introduce several innovations to make diffusion models suitable for sequential environments; designing suitable architectures, investigating the role of guidance, and developing reliable sampling strategies.

Experimentally, diffusion models closely match human demonstrations in a simulated robotic control task and a modern 3D gaming environment.