BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Text To Programs: Automating Computer Tasks from Natural Language 
 Descriptions - Nate Kushman\, MIT
DTSTART:20150309T090000Z
DTEND:20150309T100000Z
UID:TALK58373@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:As computers have become more powerful and more sophisticated\
 , they have also become more complicated and difficult for humans to use. 
  Many tasks that a user would like to perform\, however\, have already bee
 n described on-line in natural language.  By designing new machine learnin
 g algorithms that can translate these natural language descriptions into c
 omputer programs\, we enable machines to perform many tasks which today re
 quire significant manual effort.\n\nIn this talk I will describe systems I
  have built towards this end in three different computing domains:  text p
 rocessing tasks\, mathematical calculations\, and robotic object construct
 ion in a virtual world.  By translating the natural language to computer p
 rograms we are able to both determine semantic equivalence between the pro
 grams\, and observe the results of their execution.  I will show that inte
 grating these capabilities into the learning algorithms significantly impr
 oves the effectiveness of the learning\, as well as enabling learning with
  much less training data.\n
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station Road\, Cambridge
 \, CB1 2FB
END:VEVENT
END:VCALENDAR