My name is Pierre Lison, and I am a Senior Research Scientist at Norsk Regnesentral (Norwegian Computing Centre, often shortened as NR), a contract-funded research institute located in Oslo, Norway.
My research interests are in natural language processing (language technology), statistical modelling and machine learning. In the past, I have been involved in various projects related to human-robot interaction, conversational interfaces and statistical machine translation.
One of my current projects (funded by the Norwegian Research Council) investigates how to improve machine translation technology for conversational domains, such as film subtitles. In spite of considerable progress in recent years, machine translation systems remain poor at adapting translations to the context. We are therefore developing novel, “context-aware” translation models with the hope of improving translation quality. I also worked for many years on the development of spoken dialogue systems, and am the main developer behind the OpenDial toolkit which is used to build robust and adaptive dialogue systems.
Since I joined NR, I’m also involved in the centre for research-based innovation Big Insight: Statistics for the knowledge economy.
Feel free to contact me if you have any questions or ideas you would like to discuss!
I’m thrilled to start a new position as a senior research scientist at Norsk Regnesentral and am looking forward to working on new exciting projects!
Jörg Tiedemann and I just released a new major release of the OpenSubtitles collection of parallel corpora. The release is compiled from a large database of movie and TV subtitles and includes 2.6 billion sentences across 60 languages! This makes it the world’s largest collection of parallel corpora freely available. See our paper at LREC for more details on the corpus.
My article “A hybrid approach to dialogue management based on probabilistic rules” will soon appear in Computer Speech & Language (see the “Publications” section). The article is essentially a summary of my PhD work and describes the formalisation of probabilistic rules, the statistical estimation of unknown parameters, and the empirical evaluation of the framework in a human-robot interaction domain.