TCC Division    
 Cognitive and Communication
Via Sommarive, 18 I-38050
Povo-Trento, Italy
 Technologies
  You Are Here: HomeResearch Areas

Research Areas
   


Textec

Current research in Text Processing Technologies (Textec) at ITC-irst includes Question/Answering (we are participating at TREC-2001 with the DIOGENE system), Word Sense Disambiguation (we participated at SENSEVAL-II) and Information Extraction. We are involved in the construction and maintenance of MultiWordNet, a multilingual (English/Italian) lexical database based on the WordNet model. Machine Translation, currently speech-to-speech translation, is also investigated within the NESPOLE! EU project.

The general aim is to develop tools for processing texts at different levels of complexity, including morpho-syntactic analysis and lexical-semantic relations. The approach is based on the integration of knowledge based and statistical techniques.

Application areas we are interested in include: information extraction from text in specialized domains, open domain question answering on the Web, cross lingual information retrieval, text categorization and speech to speech translation in a call center context.

     


i3p

The i3p Project (Intelligent Interactive Information Presentation) addresses the problem of automatic presentation of information.

The impressive amount of information available today through Internet and electronic publishing demands new functionalities for users to be able to access it in a "natural" and flexible way. While mobile information access is already possible today, new challenges for tomorrow will be in personalized and context-sensitive information presentation. Of particular importance will be proactive devices that provide new information not explicitly requested and that stimulate new interests in the user.

The main goal of the i3p Project is to investigate innovative technologies for information presentation with a particular interest on the requirements imposed by new devices like PDAs (Personal Digital Assistants), third generation mobile phones and e-books. Given the background of the TCC division, we are mainly committed to those scenarios and technologies where natural language plays an important role while also considering interaction between natural language and other modalities.

The main research issues are (i) automatic generation of multimedia presentation adapted to the specific context (i.e. the user model, the interaction history, the physical location, the device, etc.) and (ii) the study of interactive interfaces based on the intelligent agents' metaphor (instead of being based on the navigation metaphor, like the web).
The applicative scenarios include multimedia presentation of information for cultural heritage (e.g. museums and historical towns) and edutainment, both on mobile and fixed platforms.