The Universität des Saarlandes (UdS, www.uni-saarland.de) and the German Research Center for Artificial Intelligence (DFKI, www.dfki.de) are opening a post-doctoral position (post-doc) in:
“Neural Feature and Representation Learning for Translation and Translation Technology”.
The position is funded by the Collaborative Research Cluster (CRC) "Information Density and Linguistic Encoding" (SFB 1102 www.sfb1102.uni-saarland.de/) Project B6 at UdS and the Deeplee project (https://www.deeplee.de) at DFKI. The projects are complementary in methodology and objectives and are lead by the same PIs. Successful applicants will be employed both at UdS and DFKI, with separate contracts.
Responsibilities: fundamental research, publication of research outcomes, ML software development, contribution to supervision and teaching
Requirements: PhD in language technology, computer science, machine learning (or similar); strong background/publications in MT/NLP, machine learning, deep learning; strong problem solving and programming skills; independent and creative thinking; strong team working and communication skills; excellent command of written and oral English.
Command of German and/or other languages helpful, but not a requirement.
Application deadline: 28th February 2019
Start date: as soon as possible
Location: UdS and DFKI, Campus Saarland University, Saarbrücken, Germany
Duration: 2 years.
SFB B6: translated text shows characteristics that distinguish it from comparable text originaly authored in the target language. These characteristics are often refered to as "transationese". Machines (ML) are good at distinguishing originally authored from translated texts. To date research has mostly focused on traditional human feature-engineering based supervised ML for translationese classification. Research objectives: investigate (i) the performance of deep learning based representation learning, (ii) whether the features/representations learned support linguistic and/or information-theoretic interpretations (e.g. Shannon suprisal, information density), (iii) whether insights obtained can improve (N)MT.
DFKI Deeplee: given enought data, neural approaches often outperform alterantive approaches in NLP. Research objectives: (i) carry out foundational research in neural machine translation (NMT), including (ii) NMT architectures, (iii) use of data, (iv) inclusion of (external) knowledge sources and (iv) explainability of models.
Successful applicants will work in the Collaborative Research Cluster "Information Density and Linguistic Encoding" in the Language Science and Technology (LST) Department at the UdS, and the Multilingual Technologies (MLT) Lab at DFKI. Both LST and MLT are leading centres for language technology research and provide dynamic and stimulating international research environments.
The UdS Campus hosts top-ranked collaborating research institutions including the German Research Centre for Artificial Intelligence (DFKI), the Max Planck Institute for Informatics (MPI-INF), the Max Planck Institute for Software Systems (MPI-SWS), the Helmholtz Center for Computer Security (CISPA), the Center for Bioinformatics (CBI) and the Computer Science Department.
Applications should include: short cover letter, CV, list of publications, brief summary of research interests, contact information for two references.
Please send your electronic application (PDF) to Dr. Raphael Rubino (raphael.rubino at dfki.de) and Prof. Josef van Genabith (josef.van_genabith at dfki.de) referring to the position (post-doc SFB/DFKI). The position remains open until filled.