Hi all,
Join us for our next seminar featuring Michael Pradel from the University of Stuttgart. Register here to receive the Zoom link.
Title: Neural Software Analysis: Recent Advances on Types, Bugs, and Executions
Abstract: Neural software analysis is an emerging approach for addressing difficult program analysis problems by exploiting the predictive power of deep neural networks. This talk gives an overview of our recent advances in neural software analysis, focusing on three kinds of problems. First, we present a neural program repair technique that fixes static type errors in Python. Second, we present work on neural bug detection, which detect incorrect code by predicting that two statements are inconsistent with each other. Finally, we present work on learning-guided execution, which enables the execution of incomplete code snippets by filling in missing information via queries to a neural model.
The seminar will take place on Monday, June 26, at 1 PM ET (click here for other time zones).
We look forward to having you.
Best,
Nadav