BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Learning to Represent Programs with Graphs
DTSTART:20180126T140000
DTEND:20180126T150000
DTSTAMP:20260407T043841Z
UID:d02477ec9cbeffecaac6766d79860a25cf55f09f20c5ef8af72a29c5
CATEGORIES:Conferences - Seminars
DESCRIPTION:By Miltos Allamanis - Microsoft\n\nAbstract\nThe abundance of 
 available source code raises the exciting possibility to develop novel mac
 hine learning-based software engineering tools that learn from large corpo
 ra. These probabilistic source code analyses necessitate interdisciplinary
  research in machine learning and programming languages. The core challeng
 e rests in developing methods that can probabilistically handle the multif
 aceted and highly structured aspects of source code.\nIn the talk\, I disc
 uss a graph-based neural network that learns to exploit the rich syntactic
  and semantic structure of code to detect variable misuse bugs. This model
  learns distributed representations of programs from a corpus of 2.9 billi
 on lines of code and can correctly reason about the correct variable usage
  82% of the times. This model has already detected previously unknown bugs
  in open-source projects.\n\nBio\nI am a researcher at Microsoft Research 
 in Cambridge\, UK. My research concerns application of machine learning an
 d natural language processing to software engineering and programming lang
 uages to create smart software engineering tools for developers. In the er
 a of “big code”\, code is a form of data that can be manipulated by ma
 chine learning methods to provide useful software engineering tools\, inte
 rfaces and insights. I focus on developer tools with a strong machine lear
 ning component\, while using problems of this area to motivate machine lea
 rning research. Previously\, I was a postdoc in Microsoft Research and obt
 ained my PhD in the University of Edinburgh advised by Dr. Charles Sutton.
 \n\nMore information
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR