Bug hunting with machine learning

That the software is defective and needs endless patches during its life cycle, is something that you and I always know and something that I already told in the post that opened this blog, now it may be that thanks to machine learning we are one more step close to getting more robust applications, creating systems capable of generating patches more quickly, and much more efficiently, learning to identify bugs and how to solve them.

This is something that the Program Analysis and Compilation group of MIT CSAIL (Computer Science and Artificial Intelligence Laboratory) is already working on in its Patch Generation project via Learning and how it can be read in its paper Automatic Inference of Code Transforms and Search Spaces for Automatic Patch Generation Systems.

Undoubtedly, this in my opinion can be one of the applications of machine learning with greater impact, since nowadays, our whole world is in the code, and a robust and efficient code, ensures that our future does not fall apart.