Giuseppe Carenini, University of British Columbia
Assistant Professor
Computer Science
Vancouver, British Columbia
carenini@cs.ubc.ca
Office:
(604) 822-5102
Expert In
Bio/Research
Innumerable studies indicate that natural human communication is so effective because it is adaptive, incremental, and multimedia. If we want to develop interactive computer systems which aim to match (and eventually to surpass) the effectiveness of human-human communication, they must be endowed...
Click to Expand >>
Click to Expand >>
Bio/Research
Innumerable studies indicate that natural human communication is so effective because it is adaptive, incremental, and multimedia. If we want to develop interactive computer systems which aim to match (and eventually to surpass) the effectiveness of human-human communication, they must be endowed with the same properties.
My research goal is to enable computers generate effectiveexplanations, arguments, reports, summaries and narratives that: (i) are tailored to the user and the communicative context (ii) effectively combine natural language and information graphics (iii) provide convenient, interactive means for enabling users to further explore the information presented.
I believe that devising flexible generation systems must rely on integrating and extending principles and techniques from several disciplines including computational linguistics, user modeling and information visualization.
I consider empirical evaluation a crucial step in my research. The aim is to develop generation systems that can be effectively tested. Since the evaluation of user and context tailored, interactive presentations tipically requires an explicit reference to a user task, additional research challenges must be faced in devising tasks that allow the assessment of a presentation effectiveness in controlled experiments.
Click to Shrink <<
My research goal is to enable computers generate effectiveexplanations, arguments, reports, summaries and narratives that: (i) are tailored to the user and the communicative context (ii) effectively combine natural language and information graphics (iii) provide convenient, interactive means for enabling users to further explore the information presented.
I believe that devising flexible generation systems must rely on integrating and extending principles and techniques from several disciplines including computational linguistics, user modeling and information visualization.
I consider empirical evaluation a crucial step in my research. The aim is to develop generation systems that can be effectively tested. Since the evaluation of user and context tailored, interactive presentations tipically requires an explicit reference to a user task, additional research challenges must be faced in devising tasks that allow the assessment of a presentation effectiveness in controlled experiments.
Click to Shrink <<