Björn Gambäck was appointed professor in language technology at IDI in November 2007. He will be sharing his time evenly between IDI and his work as Research Expert at the Swedish Institute of Computer Science (SICS).
At SICS he is working as Principal Investigator in the European Commission's 6th Framework Programme Integrated Project COMPANIONS on embodied conversational agents (ECAs). The project concerns ECAs as persistent companions for users in carrying out activities within their personal digital space, such as organising image and text records of their lives as coherent narratives.
Björn's research career started in September 1989 at the SICS, writing a thesis on pragmatic reasoning in Bridge as part of the MSc (civ.ing.) programme in Computer Science and Engineering (Datateknik) at KTH / The Royal Institute of Technology.
His research work in the early 90s was centred around building unification-based syntactic grammar, morphology and lexica for Swedish, machine translation, and machine learning for lexical acquisition - work that lead up to the writing of his PhD thesis.
In July 1995 was offered a position at the Research Group Head level ("BAT 1a") at the University of the Saarland, Saarbrücken, Germany. The visit to Germany changed his research focus towards computational and robust semantics for underspecified representations and to spoken dialogue understanding for German and Japanese (within the Verbmobil spoken-dialogue machine translation project).
In 1997 Björn returned to Sweden and the speech group at KTH where he defended his PhD thesis. Afterwards he was offered a position as full professor (Chair) at the University of Helsinki, Finland substituting half-time for Kimmo Koskenniemi for two years. From 1997 to 1999 he worked both in Helsinki and in Stockholm, first at KTH and from 1998 at SICS again.
Since 2001 he has been coordinating projects sponsored by the European Commission. First DUMAS, an effort within the 5th Framework Programme aimed at building adaptive multilingual mobile spoken dialogue systems using robust language understanding and machine learning-based user adaptation. Then the 6th Framework Programme (FET Open) Integrated Project EVERGROW on complex systems. The key points of the project concerned the evolution of cooperative behaviour in networks of vast numbers of independent agents, each playing to its self-interest and to the limits of its own knowledge.
In spring of 2004 he visited the Addis Ababa University in Ethiopia and has since remotely supervised a group of master students who worked on Self-Organizing Maps for text classification and Speech Recognition for Amharic (the main language of Ethiopia).
Since then he has worked part time on a project to use machine learning methods for rapid development of language processing tools for Amharic, in a project called Language Processing Resources for Under-Resourced Languages sponsored by Sida through SPIDER, the Swedish Programme for ICT in Developing Regions.
Source: Björn's homepage.