Learning to Visually Navigate in Photorealistic Environments Without any Supervision
It addresses the challenge of autonomous navigation for agents in realistic settings where position information is unavailable, representing an incremental advance in unsupervised learning methods.
The paper tackles the problem of learning visual navigation in photorealistic environments without external supervision, achieving successful navigation in challenging Gibson dataset settings using only RGB inputs.
Learning to navigate in a realistic setting where an agent must rely solely on visual inputs is a challenging task, in part because the lack of position information makes it difficult to provide supervision during training. In this paper, we introduce a novel approach for learning to navigate from image inputs without external supervision or reward. Our approach consists of three stages: learning a good representation of first-person views, then learning to explore using memory, and finally learning to navigate by setting its own goals. The model is trained with intrinsic rewards only so that it can be applied to any environment with image observations. We show the benefits of our approach by training an agent to navigate challenging photo-realistic environments from the Gibson dataset with RGB inputs only.