Learning to Navigate Using Mid-Level Visual Priors

Abstract

How much does having visual priors about the world (e.g. the fact that the world is 3D) assist in learning to perform downstream motor tasks (e.g. navigating a complex environment)? What are the consequences of not utilizing such visual priors in learning? We study these questions by integrating a generic perceptual skill set (a distance estimator, an edge detector, etc.) within a reinforcement learning framework (see Fig. 1). This skill set ("mid-level vision") provides the policy with a more processed state of the world compared to raw images. Our large-scale study demonstrates that using mid-level vision results in policies that learn faster, generalize better, and achieve higher final performance, when compared to learning from scratch and/or using state-of-the-art visual and non-visual representation learning methods. We show that conventional computer vision objectives are particularly effective in this regard and can be conveniently integrated into reinforcement learning frameworks. Finally, we found that no single visual representation was universally useful for all downstream tasks, hence we computationally derive a task-agnostic set of representations optimized to support arbitrary downstream tasks.

In Conference on Robot Learning, 2019. See project website and demos at http://perceptual.actor/

References in corpus (12)

• Deep Residual Learning for Image Recognition
• High-Dimensional Continuous Control Using Generalized Advantage Estimation
• End-to-End Training of Deep Visuomotor Policies
• OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
• On Evaluation of Embodied Navigation Agents
• Dense Object Nets: Learning Dense Visual Object Descriptors By and For Robotic Manipulation
• Learning to Poke by Poking: Experiential Learning of Intuitive Physics
• DARLA: Improving Zero-Shot Transfer in Reinforcement Learning
• Learning to Act by Predicting the Future
• Visual Memory for Robust Path Following
• S-RL Toolbox: Environments, Datasets and Evaluation Metrics for State Representation Learning
• RGB-Depth SLAM Review

Cited by in corpus (5)

• Multi-Task Reinforcement Learning with Soft Modularization
• Robust Policies via Mid-Level Visual Representations: An Experimental Study in Manipulation and Navigation
• Reinforced Wasserstein Training for Severity-Aware Semantic Segmentation in Autonomous Driving
• Unsupervised Domain Adaptation for Visual Navigation
• DeepMPCVS: Deep Model Predictive Control for Visual Servoing