BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Deep learning models for dependency parsing and retrieval tasks
DTSTART:20190828T140000
DTEND:20190828T160000
DTSTAMP:20260508T033010Z
UID:a47e452c59b4e190cf3e57ce5380c0325783e41a27493c3916a02f9f
CATEGORIES:Conferences - Seminars
DESCRIPTION:Alireza Mohammadshahi \nEDIC candidacy exam\nExam president: 
 Prof. Martin Jaggi\nThesis advisor: Prof. Karl Aberer\nThesis co-advisor: 
 Prof. James Henderson\nCo-examiner: Dr. Martin Rajman\n\nAbstract\nNatural
  language processing (NLP) is one of the most important technologies of th
 e information age. Understanding complex language utterances is also a cru
 cial part of artificial intelligence.  There is a wide range of  NLP app
 lications such as machine translation\, question answering\,  parsing\, r
 etrieval task\, etc. In recent years\, the striking success of deep learni
 ng in a wide variety of natural language processing (NLP) applications has
  served as a new benchmark for future advances. In this report\, I especia
 lly concentrate on two critical applications of NLP\, named dependency par
 sing and image-caption retrieval task\, and I describe conventional models
  for these challenging tasks.\n\nBackground papers\nTransition-Based Depen
 dency Parsing with Stack Long Short-Term Memory\, by Dyer\, C.\, et al.\nI
 mage Pivoting for Learning Multilingual Multimodal Representations\, by Ge
 lla S.\, et al.\nGraph-based Dependency Parsing with Bidirectional LSTM\, 
 by Wang\, W.\, Chang\, B.\n 
LOCATION:INF 211 https://plan.epfl.ch/?room==INF%20211
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR