Post-doc – Machine Learning and Optimization group – language models as knowledge bases

NAVER LABS Europe is the biggest industrial AI research lab in France. As the research arm of Korea's leading internet services company, NAVER, the mission of LABS is to create future technologies and new ways to interact and connect to the digital and real world. This includes research in neural systems to give humans faster and better access to information and knowledge.

To achieve this, we need to dramatically improve information access and question answering. Over the last few years, the paradigm shift of deep learning and self-attention networks in data science has dramatically improved performance over the full spectrum of applications and, optimized through self-supervised encoding tasks, these networks have allowed for the building of both syntactic and semantic a priori knowledge for natural language understanding and generation tasks. Besides these representation capabilities, it has been shown that pre-trained, large-scale parametric language models capture, as a side effect, part of the information enclosed in texts used for their training. As such, they tend to keep question-answering capabilities using cloze-form statements or open-domain formulations for masked or causal language model approaches respectively.

In this context, NAVER LABS Europe is looking for a motivated, post-doctoral researcher to develop novel modelisation and learning approaches to address the problem of handling compositional questions with possible applications to NAVER Question-Answering services.

Parametric information storage and retrieval has been studied in other contexts such as in images, and constitutes a promising research direction for fundamental machine learning. During this post-doctoral study, the successful candidate will first conduct an exhaustive evaluation of the capabilities of parametric language models as a knowledge base before developing novel modelisation, learning and adaption paradigms to enhance the usability of such models in challenging question-answering settings.

- Ph.D. in spoken language processing, speech processing, NLP or machine learning.
- Knowledge of the latest developments in statistical and deep learning as applied to NLP and speech.
- Strong publication record in top-tier NLP, speech or machine learning conferences.
- Strong development skills, preferably in Python and knowledge of relevant frameworks (TensorFlow, PyTorch, etc).
- Fluent in Korean language (spoken and written) as involves working on and understanding Korean data.

You can apply for this position online. Don't forget to upload your CV and cover letter before you submit.
Incomplete applications will not be accepted.

Due to the changing travel restrictions related to COVID-19, it may not be possible to host candidates from certain regions. This will depend on the conditions at the specific starting date of the internship.