In my presentation I will talk about how I broaden sentiment analysis in order to equip computers with knowledge needed to perform ethical decisions. While researchers of the young field of Machine Ethics tend to utilize normative ethics (telling how things should be), I decided to approach the problem from the comparative ethics' (saying how things are) point of view. By adding social consequences analysis to the emotive polarization I try to extend the existing text-mining techniques and perform experiments with one step deeper analysis of context which is crucial in recognizing if an author expressed a negative or positive statement. The latest results and considerations will be discussed. I will also briefly introduce systems under development at his interdisciplinary Language Media Laboratory at Hokkaido University, Japan (affect recognition, language acquisition, dialog processing, humor generation, etc.).
Last modified: Wednesday, 13-Mar-2013 12:22:01 NZDT
This page is maintained by the seminar list administrator.