Language and Interaction Technologies Group

Language Technology

has been a major research area at the Austrian Research Institute for Artificial Intelligence (OFAI) since its foundation in 1984. In this area, we conduct basic and applied research in modelling and processing human languages at text and speech levels. This includes the construction of
  • linguistic resources -- such as general language and special purpose lexicons, relational databases, RDF stores, and ontologies
  • natural language processing components -- such as tokenizers, morpho-syntactic processors, semantic analysers, sentiment detectors, text generators and summarization as well as dialogue components
  • text-to-speech systems -- such as speech synthesisers for Austrian German and Viennese varieties
  • full application prototypes -- such as
    • netiquette assessment for forum postings in online newspapers
    • targeted expert search for consulting companies
    • semantic clustering for free text associations in online brainstormings, answers in opinion polls, and targeted interviews
    • computational linguistics enhanced tools for the Digital Humanities -- such as corpus annotation tools for joint qualitative and quantitative analyses of textual data

Interaction Technology

is another major research area at OFAI. Respective research started at OFAI in 2000 and is concerned with the modelling of complex artificial systems that communicate and interact with the human user in a most natural way. Thus taking into account and integrating signals from different communication channels such as text, gesture, body and speech, as well as the visual and auditory context in communicative situations. Our work in this area includes
  • at macro level -- the analysis and generation of multimodal communicative behaviour in embodied conversational characters and robots
  • at micro level -- research and development of methods and tools for
    • multimodal processing of natural language utterances
    • question answering and conversational systems
    • sentiment classification and (text-based) mining for opinions and affective states
    • multimodal natural language generation
    • development of representation formats for multimodal behaviours
    • creation of multimodal data sets
    • multimodal language learning for artificial agents