پوزيشن دکتری دانشگاه گرونوبل فرانسه

——————————————
✅PhD Position in Deep Learning at LIG (Grenoble): Coreference Resolution and Machine Translation

 

✅Ph.D. proposal for the project ANR JCJC CREMA

✅Coreference REsolution into MAchine Translation (CREMA)

✅The CREMA project aims at improving Document-Level Neural Machine
Translation models (DL-NMT) by integrating a module for coreference
resolution in a neural translation model. DL-NMT is currently one of
the most interesting research directions, and it has remarkable fallout
both on scientific and applicative domains.
Current models for DL-NMT integrate a context for improving translation
consisting of a fixed number of previous sentences. Previous work has
shown however that most of the words in a sentence can be correctly
translated without any context. While words needing a context are
relatively rare in sentences, their correct translation has a
remarkable impact on the quality of document translations. Moreover,
information contained in the context and needed for the correct
translation of some words are rare as well.
As consequence, the utilisation of a fixed number of sentences as
context may drawn in the noise words bringing such information. The
literature has shown that the most ambiguous words to be translated are
those involved in discourse phenomena, in particular anaphora and
coreferences. The objectif of the project CREMA is thus to integrate a
coreference resolution module in a neural translation system. This
module will detect sentences containing words needing a context for
their translation, as well as those sentences containing disambiguating
clues. The translation model will be able thus to use only such
sentences as context, instead of a fixed number of sentences chosen in
advance.

The Ph.D. program will focus at first on creating an end-to-end
coreference resolution system. Such system will be inspired from
existing systems and will be evaluated on the same data. In a second
time, the Ph.D. objective will be to integrate the coreference
resolution model into an existing neural translation system. The
complete system will be evaluated on the traditional data sets used for
the evaluation of DL-NMT models.

❗️Candidate profile:
     –   Strong experience in programming and machine learning for
         Natural Language Processing (NLP), particularly in deep
         learning
     –   Master with a component on NLP or computational linguistics
     –   Good knowledge of the French and/or English language

Practical details:
     –   Ph.D. begin is between January and March 2022
     –   Full time doctoral contract at the LIG laboratory
         (Getalp group) for 3 years (raw salary: 1768€ per month)
————————————

علاقه مندان جهت اطلاعات بيشتر از نحوه ارتباط و اپلاي اين پوزيشن و مشاوره، با ادمين كانال در ارتباط باشند

————————————
#پوزيشن_دكتري #دكتري_فرانسه #دانشگاه_گرونوبل #ماشين_لرنينگ #كامپيوتر #برنامه_نويسي #موقعيت_دكتري_با_فاند
#phd #phd_position #France #Grenoble #mashine_learning