Portage: A high-performing SMT system that is pure phrase-based (almost)

======================================================================
                Joint Seminar
======================================================================
The Hong Kong University of Science & Technology
Human Language Technology Center
Department of Computer Science and Engineering
Department of Electronic and Computer Engineering
---------------------------------------------------------------------

Speaker:        Dr. Roland Kuhn
                Principal Research Officer
                The National Research Council of Canada (NRC)

Title:          "Portage: A high-performing SMT system that is pure
                 phrase-based (almost)"

Date:           Wednesday, 11 December 2013

Time:           4::00pm - 5:00pm

Venue:          Lecture Theater F (near lifts 25 & 26), HKUST

Abstract:

Beginning with a brief reminder of the main characteristics of
phrase-based statistical machine translation (SMT), the talk will describe
NRC's Portage system. It will focus on the version that participated in
the NIST 2012 Arabic-English and Chinese-English MT evaluation. The talk
will seek to convey the excitement of developing a system that will
compete with other world-class systems in a major evaluation. Four
techniques mainly responsible for Portage's success in the NIST 2012
evaluation will be described:

-       batch lattice MIRA
-       discriminative hierarchical reordering
-       multiple phrase pair extraction
-       domain adaptation with linear mixtures

The talk will also discuss post-2012 research in SMT by the NRC group.

Finally, if there is time, the talk will give a short overview of recent
work at NRC on word-emotion associations.


*************************
Biography:

Roland Kuhn is a Principal Research Officer with the National Research
Council of Canada (NRC). After obtaining his PhD in Computer Science from
McGill University in 1993, he worked for the Centre de recherche
informatique de Montréal (CRIM) until September 1996. Subsequently, he
worked for Panasonic Speech Technology Laboratory in Santa Barbara,
California (Oct. 1996 - June 2004). During this first period of his
career, his research focused on areas related to speech: automatic speech
recognition, speaker adaptation, dialogue, and speaker
verification/identification. In July 2004, he joined NRC and embarked on
research in machine translation; he is co-founder and head of NRC's
Portage machine translation team. He has authored 61 refereed publications
and holds 29 US patents.