The (Un)Surprising Effectiveness of Pre-Trained Vision Models for Control
Simone Parisi∗ Aravind Rajeswaran∗ Senthil Purushwalkam Abhinav Gupta
Abstract
Recent years have seen the emergence of pre-trained representations as a powerful abstraction for AI applications in computer vision, natural language, and speech. However, policy learning for control is still dominated by a tabula-rasa learning paradigm, with visuo-motor policies often trained from scratch using data from deployment environments. In this context, we revisit and study the role of pre-trained visual representations for control, and in particular representations trained on large-scale computer vision datasets. Through extensive empirical evaluation in diverse control domains (Habitat, DeepMind Control, Adroit, Franka Kitchen), we isolate and study the importance of different representation training methods, data augmentations, and feature hierarchies. Overall, we find that pre-trained visual representations can be competitive or even better than ground-truth state representations to train control policies. This is in spite of using only out-of-domain data from standard vision datasets, without any in-domain data from the deployment environments. Source code and more at https://sites.google.com/view/pvr-control.
中文速览
用预训练好的视觉模型（Pre-trained Visual Representations, PVR）直接驱动机器人控制策略，一直被认为效果有限，因为控制任务与普通视觉任务差异显著。研究者系统评估了多种主流视觉预训练模型（包括有监督的ResNet、自监督的MoCo以及CLIP）在四个控制仿真平台（Habitat导航、DeepMind Control、Adroit灵巧手操控、Franka Kitchen）上的表现，这些模型从未见过任何部署环境中的画面。结果出人意料：以自监督方式训练的视觉模型，配合来自模型不同层级的特征融合，竟能媲美甚至超越直接使用模拟器提供的"真实状态"特征来训练策略；进一步分析发现，自监督学习优于有监督学习，裁剪数据增强比颜色增强更关键，浅层特征适合精细运动控制而深层特征更适合语义导航任务。这项工作表明，不依赖部署环境数据、仅凭通用大规模视觉预训练就能为控制策略提供高质量感知模块，为低数据、高泛化的机器人学习指出了一条切实可行的路径。
原文 arXiv:2203.03580;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2203.03580v2