BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Colloquium : Statistical Analysis of Computer Program Text: Mac
 hine Learning and Natural Language Processing Meets Software Engineering
DTSTART:20151109T161500
DTEND:20151109T173000
DTSTAMP:20260407T051146Z
UID:dc355fd5023de78a92a20dbb3fab5a3cfce54256790ced712772524f
CATEGORIES:Conferences - Seminars
DESCRIPTION:By : Charles Sutton - University of EdinburghVideo of his talk
 Abstract :\nBillions of lines of source code have been written\, many of w
 hich are freely available on the Internet. This code contains a wealth of 
 implicit knowledge about how to write software that is easy to read\, avoi
 ds common bugs\, and uses popular libraries effectively.\nWe want to extra
 ct this implicit knowledge by analyzing source code text.\nTo do this\, we
  employ the same tools from machine learning and natural language processi
 ng that have been applied successfully to natural language text.\nAfter al
 l\, source code is also a means of human communication.\nWe present three 
 new software engineering tools inspired by this insight:\n* Naturalize\, a
  system that learns local coding conventions.\nIt proposes revisions to na
 mes and to formatting so as to make code more consistent.\nA version that 
 uses word embeddings has shown promise toward naming methods and classes.\
 n* Data mining methods have been widely applied to summarize the patterns 
 about how programmers invoke libraries and APIs. We present a new method f
 or mining market basket data\, based on a simple generative probabilistic 
 model\, that resolves fundamental statistical pathologies that lurk in pop
 ular current data mining techniques.\n* HAGGIS\, a system that learns loca
 l recurring syntactic patterns\, which we call idioms. HAGGIS accomplishes
  this using a nonparametric Bayesian tree substitution grammar\, and is de
 licious with whisky sauce.Bio :\nCharles Sutton is a Reader (equivalent to
  Associate Professor: http://bit.ly/1W9UhqT) at the University of Edinburg
 h. He is interested in a broad range of applications of probabilistic mach
 ine learning\, including NLP\, analysis of computer systems\, software eng
 ineering\, sustainable energy\, and exploratory data analysis.\nDr Sutton 
 completed his PhD at the University of Massachusetts Amherst\, working wit
 h Andrew McCallum. He did postdoctoral research at the University of Calif
 ornia Berkeley\, working with Michael I Jordan.\nHe is Deputy Director of 
 the EPSRC Centre for Doctoral Training in Data Science at the University o
 f Edinburgh.More information
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR