NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:5389
Title:Devign: Effective Vulnerability Identification by Learning Comprehensive Program Semantics via Graph Neural Networks

The paper describes a technique for identifying functions in code that may suffer vulnerabilities. The problem is important and timely. The technique is evaluated on real data, and the results seem very promising. The reviewers liked the work and appreciated the response made by the authors.