My name is Pierre Lison, and I am a Senior Research Scientist at Norsk Regnesentral (Norwegian Computing Center, NR for short), an independent research institute located in Oslo, Norway. I am also a part-time Associate Professor at the Language Technology Group, University of Oslo.

My research interests are in natural language processing (language technology), statistical modelling and machine learning. In my previous research, I worked on topics such as human-robot interaction, conversational interfaces and statistical/neural machine translation.

I am currently leading two research projects, both funded by the Research Council of Norway. The first, CLEANUP, seeks to develop new, data-driven methods for removing personal information from text data. The other, GraphDial, focuses on dialogue management and investigates the use of graphs (and weak supervision) to represent the dialogue state of rich conversational domains.

Other research projects I have been involved with include SAFERS (speech analytics for emergency response services), DialMT (dialogue modelling for machine translation) and Oslo Analytics (cyber-threat intelligence). I also participated in the centre for research-based innovation Big Insight: Statistics for the knowledge economy.

I’m also a member of the Young Academy of Norway.



We have just released skweak, a new Python library to facilitate the use of weak supervision in NLP tasks. For more info, check out our repo!


Glad to have recruited two fantastic PhD students: Anthi Papadopoulou works on new models for automated text anonymisation as part of the CLEANUP project, while Nick Walker will focus on graph-based dialogue state tracking as part of the GraphDial project. Welcome onboard!


Ingrid Lossius Falkum and myself wrote an article to debunk some of the hype around neural language models in Aftenposten (Norway's second largest newspaper).