Wield: Systematic Reinforcement Learning With Progressive Randomization

09/15/2019
by   Michael Schaarschmidt, et al.
0

Reinforcement learning frameworks have introduced abstractions to implement and execute algorithms at scale. They assume standardized simulator interfaces but are not concerned with identifying suitable task representations. We present Wield, a first-of-its kind system to facilitate task design for practical reinforcement learning. Through software primitives, Wield enables practitioners to decouple system-interface and deployment-specific configuration from state and action design. To guide experimentation, Wield further introduces a novel task design protocol and classification scheme centred around staged randomization to incrementally evaluate model capabilities.

READ FULL TEXT
research
11/05/2019

Gym-Ignition: Reproducible Robotic Simulations for Reinforcement Learning

In this paper we present Gym-Ignition, a new framework to create reprodu...
research
06/06/2022

Real2Sim or Sim2Real: Robotics Visual Insertion using Deep Reinforcement Learning and Real2Sim Policy Adaptation

Reinforcement learning has shown a wide usage in robotics tasks, such as...
research
12/12/2019

The PlayStation Reinforcement Learning Environment (PSXLE)

We propose a new benchmark environment for evaluating Reinforcement Lear...
research
10/23/2019

Robust Domain Randomization for Reinforcement Learning

Producing agents that can generalize to a wide range of environments is ...
research
11/13/2020

DeepMind Lab2D

We present DeepMind Lab2D, a scalable environment simulator for artifici...
research
10/03/2019

Running Alchemist on Cray XC and CS Series Supercomputers: Dask and PySpark Interfaces, Deployment Options, and Data Transfer Times

Newly developed interfaces for Python, Dask, and PySpark enable the use ...

Please sign up or login with your details

Forgot password? Click here to reset