AndroidEnv provides a straightforward mechanism for flexibly creating such custom tasks based on any Android application. For example, one might set the goal of finding directions to the park, booking a flight or maximizing the score in a game.

Scale: The large size of the observation and action space poses an interesting scaling problem for RL agents.Īndroid’s large ecosystem opens up the possibility for defining varied tasks, enabling agents to learn to achieve different types of objectives on the same platform.Real-time dynamics: Services and applications run in real-time simulation, making the environment dynamics similar to robotics control tasks.Temporal abstraction: Learning gestures and flexible ways to compose actions is necessary for an agent to be able to handle the immense native action space.Transfer and generalization: The observation and action space is the same across all applications, allowing for many opportunities to transfer knowledge across tasks of very different nature.

Learning to solve tasks in AndroidEnv requires an agent to overcome multiple types of challenges that have long interested researchers: Apart from the flexibility and real-world aspect of the platform, AndroidEnv is a particularly appealing domain for RL research thanks to its diverse features.

Download game starcraft for android