Ta witryna wykorzystuje cookies. Więcej informacji można znaleźć na stronie Polityka dotycząca cookies i podobnych technologii. ZAMKNIJ Zamknij ostrzeżenie dotyczące cookies

Seminarium naukowe Zakładu Systemów Informacyjnych

zdjecie

Zakład Systemów Informacyjnych Instytutu Informatyki zaprasza na seminarium 12 listopada br. o godzinie 12:15 w Audytorium Centralnym (AC). Wykład pt. "Text analysis of patent collections" wygłosi profesor W. Zadrożny z University of North Carolina in Charlotte. Poniżej autorskie streszczenie referatu oraz krótka notka biograficzna o prelegencie.

Abstract: Patent data is the largest freely available collection of technical text. So far it hasn’t been widely explored. In this talk I will describe our  recent work on classification of patents and on  patents as an early indicator of emerging technologies. For patent classification, we show 95% reduction in number of linguistic features, with no decrease in accuracy of classification, by using features derived from Wikipedia titles and Google bigrams. In viewing patent as an indicator of innovation,  we show that at least some of the emerging technologies (smart phones) could have been discovered prior official announcements by analyzing granted patents and patent application. We’ll also describe some of the ontology mapping problems arising in analysis of patent data.

Speaker’s Bio. Wlodek Zadrozny is Associate Professor of Computer Science at University of North Carolina in Charlotte. He received his PhD in Mathematics from Polish Academy of Science in 1980. From 1985 to his retirement in 2013 he was a researcher and manager at the IBM T.J.Watson Research Center in Yorktown Heights, NY. Dr. Zadrozny constructed first full  natural language telephony systems for banking, and created several highly successful text mining systems for industry and government. From 2008 and 2012 was responsible for textual resources for  the Watson/DeepQA project, the computer that won the TV game show Jeopardy!.  He published over fifty papers on various aspects of text processing and knowledge representation, and received over forty patents.

Ostatnia modyfikacja: środa, 26 listopada 2014, 13:02:21, Bożenna Skalska

x x Aktualności (4) - wg daty publikacji

‹‹ Listopad 2014 ››
Pon Wt Śr Czw Pt So N
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30