Using Deep Learning for Coding Tasks

Speaker:  Prof. Alvin CHEUNG
          Department of Electrical Engineering and Computer Sciences
          UC Berkeley

Title:   "Using Deep Learning for Coding Tasks"

Date:    Monday, 2 September 2024

Time:    4:00pm - 5:00pm

Venue:   Lecture Theater F
         (Leung Yat Sing Lecture Theater)
         near lift 25/26, HKUST

Abstract:

Language models (LMs) have recently been deployed in many coding-related
tasks. As many LMs are trained using a wide variety of data that spans
across code, text, and even images, they are not targeted for coding tasks
and their generated outputs come with no correctness guarantees. In this
talk, I will first discuss our recent work on using LMs for various coding
problems, from code compilation, program verification, to database query
optimization.


****************
Biography:

Alvin Cheung is an associate professor at Berkeley EECS. His group tackles
problems in programming systems and data management and is part of the
Berkeley Sky and SpeciaLIzed Computing Ecosystems (SLICE) Labs. He and his
students have received various best paper awards and distinctions, most
recently the 2024 Dahl-Nygaard Junior Prize for outstanding contribution
to programming language research, and the 2023 VLDB Early Career Award for
his group's data management work.