SZTE Innovation Award for automated bug repair

SZTE Innovation Award for automated bug repair

Viktor Csuvik's work on automatically repairing software bugs at the University of Szeged received the SZTE Innovation Award. In 2020 the same work also won the award for the most innovative student research (TDK) work. The supervisor was Dr. László Vidács.

Slide explaining the genetic algorithm
How the genetic algorithm works: a population of candidate fixes, fitness, selection, crossover and mutation.
Presenting the results
The results: fixes for 24 of 127 bugs.

The approach uses a genetic algorithm, an idea borrowed from biology. Tested on 127 real program bugs, it produced fixes for 24: 5 were identical to the fixes the developers wrote, and 7 behaved identically. A fix took about 1.5 hours on average. The starting question was simple: “Why not automate the process?”

Read the article: SZTEinfo

Planning an AI feature?

We build LLM-based products with verification, logging and human review designed in from the first sprint – and most engagements start with a free one-week proof of concept for one named process.

Book 20 minutes with ViktorSend a message