Using Reinforcement Learning to Write Java Unit Tests

Mathew Lodge ⋅ Chief Executive Officer

Simply provide your details below to access this resource:

You probably learned that hacking on code until it works is the wrong way to write a program. But that’s essentially what AI does (albeit trillions of times faster), to remarkable effect.

In this webinar, originally presented at QCon London in March 2020, Diffblue’s CEO Mathew Lodge explains how AI can use reinforcement learning to do a probabilistic search of potential solutions when the search space is polynomial—such as when playing a game of Go against a grandmaster, or when writing regression unit tests.

Learn how we apply this technique at Diffblue to generate unit test programs for Java projects, hacking the code using reinforcement learning to find the tests that achieve coverage and usefulness goals. The resulting test suites find more problems earlier in DevOps pipelines.