TL;DR
This paper introduces a novel method called Dropout's Dream Land that enhances the transfer of learned simulation environments to real-world applications by using dropout to generate diverse dream environments, improving generalization.
Contribution
It proposes a new approach to improve sim-to-real transfer by leveraging dropout during training to create varied dream environments without additional costs.
Findings
Dropout's Dream Land effectively bridges the reality gap.
The method outperforms traditional domain randomization techniques.
Extensive ablation studies validate the approach.
Abstract
A World Model is a generative model used to simulate an environment. World Models have proven capable of learning spatial and temporal representations of Reinforcement Learning environments. In some cases, a World Model offers an agent the opportunity to learn entirely inside of its own dream environment. In this work we explore improving the generalization capabilities from dream environments to real environments (Dream2Real). We present a general approach to improve a controller's ability to transfer from a neural network dream environment to reality at little additional cost. These improvements are gained by drawing on inspiration from Domain Randomization, where the basic idea is to randomize as much of a simulator as possible without fundamentally changing the task at hand. Generally, Domain Randomization assumes access to a pre-built simulator with configurable parameters but…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
MethodsDropout
