SplitNet: Sim2Sim and Task2Task Transfer for Embodied Visual Navigation
This work addresses embodied visual navigation for robotics, offering incremental advances in transfer learning and generalization.
The paper tackles the problem of visual navigation by decoupling perception and policy learning, resulting in dramatic improvements in simulator transfer and better generalization to unseen environments and tasks.
We propose SplitNet, a method for decoupling visual perception and policy learning. By incorporating auxiliary tasks and selective learning of portions of the model, we explicitly decompose the learning objectives for visual navigation into perceiving the world and acting on that perception. We show dramatic improvements over baseline models on transferring between simulators, an encouraging step towards Sim2Real. Additionally, SplitNet generalizes better to unseen environments from the same simulator and transfers faster and more effectively to novel embodied navigation tasks. Further, given only a small sample from a target domain, SplitNet can match the performance of traditional end-to-end pipelines which receive the entire dataset. Code is available https://github.com/facebookresearch/splitnet