Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This AndroidEnv code is well-written, pure-python code with no Java components [1] but still is able to do RL. Very cool.

As an Android app developer I'm wondering if this is used for better fuzzing over the classic "monkey" command [2]. In other words, can an agent learn to crash apps?

[1] https://github.com/deepmind/android_env/tree/main/android_en... [2] https://developer.android.com/studio/test/monkey



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: