DRIFT: Deep Reinforcement Learning for Functional Software Testing

Understand

Efficient software testing is essential for productive software development and reliable user experiences.

• As human testing is inefficient and expensive, automated software testing is needed.
• In this work, we propose a Reinforcement Learning (RL) framework for functional software testing named DRIFT.
• DRIFT operates on the symbolic representation of the user interface.