Fiber: A Platform for Efficient Development and Distributed Training for Reinforcement Learning and Population-Based Methods
This addresses the problem of high computational demands and adoption barriers for researchers and practitioners in RL and population-based methods, though it appears incremental as a platform improvement.
The paper tackles the challenges of efficiency and flexibility in distributed computing for reinforcement learning and population-based methods by introducing Fiber, a scalable framework that expands accessibility to large-scale parallel computation without requiring specialized expertise.
Recent advances in machine learning are consistently enabled by increasing amounts of computation. Reinforcement learning (RL) and population-based methods in particular pose unique challenges for efficiency and flexibility to the underlying distributed computing frameworks. These challenges include frequent interaction with simulations, the need for dynamic scaling, and the need for a user interface with low adoption cost and consistency across different backends. In this paper we address these challenges while still retaining development efficiency and flexibility for both research and practical applications by introducing Fiber, a scalable distributed computing framework for RL and population-based methods. Fiber aims to significantly expand the accessibility of large-scale parallel computation to users of otherwise complicated RL and population-based approaches without the need to for specialized computational expertise.