2002.
nique was applied for handling recognition errors.
Although adding information was performed using
C. Hori et. al., ”A Statistical Approach for Automatic SpeechSummarization,” EURASIP Journal on Applied Signal Pro-DQs only once, experimental results revealed the
cessing (EURASIP), pp128–139, 2003.potential of the generated DQs to acquire indispens-
Y. Sasaki et. al., “NTT’s QA Systems for NTCIR QAC-1,”able information that was lacking for extracting an-
Working Notes of the Third NTCIR Workshop Meeting,swers. In addition, the screening filter helped to gen-
pp.63–70, 2002.erate the appropriate DQs. Future research will in-
Bạn đang xem 2002. - BÁO CÁO KHOA HỌC SPOKEN INTERACTIVE ODQA SYSTEM SPIQA PDF