An introduction to automatic program repair

Feb 1, 2024·
Luke Braithwaite
Luke Braithwaite
· 0 min read
Generated by Microsoft Designer
Abstract
Deep learning has been successfully applied to detecting and repairing bugs in source code. A large amount of useful semantic information is captured inside the structure and relational properties of the source code. Graphs and hypergraphs provide a natural way to model these relationships. This talk explores how graph neural networks (GNNs) can be applies to automatic program repair (APR) tasks.
Date
Feb 1, 2024 11:00 AM — 11:45 AM
Event
Peterhouse Outreach Day
Location

Peterhouse

Trumpington Street, Cambridge, Cambridgeshire CB2 1RD

Luke Braithwaite
Authors
Luke Braithwaite
Software Engineer III, AI Deployments
I am an MPhil ACS student at Peterhouse, University of Cambridge and my research interests are graph representation learning and geometric deep learning. My current research explores adapting sheaf-based methods for heterogeneous graph data.