4-year Doctoral Student -position in language technology at the University of Helsinki
– work in project MeMAD: Methods for Managing Audiovisual Data
– PhD project: Multimodal Neural MT for audiovisual data sets
MeMAD is coordinated by Aalto University in Finland (Espoo in the Helsinki area) but the position is at the University of Helsinki in the language technology team located in the central campus of Helsinki. The consortium consists of further partners in France, the UK and Belgium. The MeMAD project runs for 3 years but this position covers 4 years.
The work of the doctoral student will be tight to the work package on machine translation and we will work in close collaboration with the team at Aalto University (experts in speech technology, video processing, unsupervised learning, data-driven NLP) and also translation studies at the University of Helsinki. Data will be provided by various partners and commercial/national content providers. The focus of the project is related to the following tasks:
– modelling multimodal neural machine translation that incorporates audiovisual information besides textual input; texts may come from automatically generated content descriptions
– development of discourse-aware machine translation: models that capture the dynamics of the story and make use of dialogue information and scene structures to improve coherence
– multilingual NMT and multi-task learning
– cross-lingual retrieval of audiovisual data based on generated descriptions and machine translation
University of Helsinki
About University of Helsinki
The University of Helsinki is one of the best multidisciplinary research universities in the world. It is an international academic community of 40,000 students and staff members, operating on four campuses in Helsinki and at 17 other locations.
Read more at www.helsinki.fi/university