BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:www.tcs.tifr.res.in/event/867
DTSTAMP:20230914T125941Z
SUMMARY:Question Answering: The Shallow and the Deep
DESCRIPTION:Speaker: Soumen Chakrabarti (Indian Institute of Technology\nDe
 partment of Computer Science\nand Engineering\nPowai\nMumbai 400076)\n\nAb
 stract: \nWeb search has come a long way from matching query words with do
 cument words.  It is now mediated by knowledge graphs (KGs) such as Freeb
 ase\, having hundreds of millions of entities belonging to tens of thousan
 ds of types\, connected by billions of relations.  Also essential is to a
 nnotate token spans in the Web corpus with canonical types (e.g. `scientis
 t') and entities (e.g. `m.0jcx'\, Freebase's unique ID for Albert Einstein
 ).  Armed with suitable indexes and ranking functions\, we can now search
  for ``scientists who played the violin''\, but only if the search engine 
 understands that `scientists' is the target type\, `violin' is a grounded 
 entity\, and `played' is the connecting relation.  We will review recent 
 dramatic improvements in the techniques search engines use to infer these 
 diverse roles of query words\, by jointly exploiting knowledge graphs and 
 corpus annotations.  The best techniques use neural networks to compare K
 G and corpus neighborhoods of candidate entities against possibly overlapp
 ing query segments.\n\nTime permitting\, I will conclude with emerging wor
 k on more complex queries such as ``countries having more rivers than Indi
 a'' or ``how old was Indira Gandhi when Rahul Gandhi was born''.  Answeri
 ng such queries involves breaking them down into subtasks and performing n
 ontrivial reasoning on the subtask responses.\n\nBio: Soumen Chakrabarti i
 s a Professor of Computer Science at IIT Bombay. He got his PhD from Unive
 rsity of California\, Berkeley and worked on Clever Web search and Focused
  Crawling at IBM Almaden Research Center. He has also worked at Carnegie-M
 ellon University and Google. He works on linking unstructured text to know
 ledge bases and exploiting these links for better search and ranking. Othe
 r interests include link formation and influence propagation in social net
 works\, and personalized proximity search in graphs. He has published exte
 nsively in WWW\, SIGKDD\, EMNLP\, VLDB\, SIGIR\, ICDE and other conference
 s. His work on keyword search in databases got the 10-year influential pap
 er award at ICDE 2012. He is also the author of one of the earliest books 
 on Web search and mining.\n
URL:https://www.tcs.tifr.res.in/web/events/867
DTSTART;TZID=Asia/Kolkata:20180424T143000
DTEND;TZID=Asia/Kolkata:20180424T153000
LOCATION:A-201 (STCS Seminar Room)
END:VEVENT
END:VCALENDAR