TU Berlin

Quality and Usability Lab2013_04_22_Osherenko

Page Content

to Navigation

There is no English translation for this web page.

Opinion mining and lexical affect sensing

*** Please notice exceptional location! ***

LOCATION: TEL1118/9, TEL, Ernst-Reuter-Platz 7, 11th floor

SPEAKER: Alexander Osherenko (Socioware Development)


This talk discusses opinion mining and lexical affect sensing. It focuses on automatic detection of emotions and opinions in texts which importance is motivated by the benefits that emotionally-intelligent technical artifacts would bring to humans. The talk presents different examples of emotional analysis such as analysis of product reviews, natural-language dialogues, weblogs and discusses an analytical (statistical) approach, a grammatical (linguistic) approach, a hybrid approach (combination of the statistical and the grammatical approaches), and a multimodal approach (combination of the lexical and the acoustic data). Since listeners of different specialties are expected to come to this talk, for instance, linguists, psychologists, computer scientists etc., we plan a discussion in a more broad context so that everybody can benefit from this talk. If, however, special questions arise in the presentation, they can be answered in the discussion. Also, a demo will be presented.


Alexander Osherenko graduated from University of Applied Sciences, Hamburg with the B.Sc. degree in computer science. He obtained his M. Sc. degree in computer science from Humboldt University, Berlin and the PhD degree from University of Augsburg. Alexander Osherenko, has successfully capitalized on the research in his thesis by founding his own company, Socioware Development. The thesis was published by SVH and is available worldwide (see a review). Recently, Alexander Osherenko applied for the grant of the Japanese Society for Promotion of Science with the proposal "Framework of analyzing social interaction in embodied conversational agents in the context of globalization".

HOST: Hamed Ketabdar


Quick Access

Schnellnavigation zur Seite über Nummerneingabe