AndroidEnv: A Reinforcement Learning Platform for Android
This provides a realistic simulation platform for RL research on Android apps, but it is incremental as it builds on existing Android and RL frameworks.
The authors introduced AndroidEnv, an open-source reinforcement learning platform built on the Android ecosystem, enabling RL agents to interact with various apps via a touchscreen interface and potentially deploy on real devices, with empirical evaluations of popular RL agents on tasks.
We introduce AndroidEnv, an open-source platform for Reinforcement Learning (RL) research built on top of the Android ecosystem. AndroidEnv allows RL agents to interact with a wide variety of apps and services commonly used by humans through a universal touchscreen interface. Since agents train on a realistic simulation of an Android device, they have the potential to be deployed on real devices. In this report, we give an overview of the environment, highlighting the significant features it provides for research, and we present an empirical evaluation of some popular reinforcement learning agents on a set of tasks built on this platform.