International Coordination of Speech Databases, Prosodic Labeling, and Speech Input/Output Systems Assessment

点击翻译按钮获取中文摘要

The aim of the project is to investigate the state of the art, and to establish international coordination in (a) speech databases, (b) prosodic labeling, and (c) assessment methods for speech recognition and speech synthesis. The activities and the results obtained during the three year period from June 1996 to March 1999 are as follows.(1) Investigation of the state of the art and discussion on international coordinationThe state of the art has been investigated in Japan as well as in U.S., U.K., France, Germany, the Netherlands, Spain, Greece, Australia, China, Korea, and Hong Kong with the cooperation of domestic and overseas members, and mechanisms and strategies were discussed for further strengthening international coordination. ((2) Establishment of international meetings as forums for exchange among researchers in related fieldsAs a result of the foregoing discussion, an international conference on language resources and evaluation has been organized in May 1998 mainly by the … More initiative of European members of the project, as a forum for presentation and discussion of related research. Considering the characteristic differences of East-Asian languages from Indo-European languages, an international workshop on language resources and evaluation of East-Asian languages has also been organized mainly by the initiative of domestic members. It was agreed that both meetings are important and will be held at regular intervals in the future.(3) In-depth discussions at workshopsThree workshops were held (Philadelphia 1996, Rhodes 1997, and Sydney 1998) for intensive discussions among members of the project as well as other researchers who are interested in international coordination. These discussions revealed the increasing importance of international coordination in this field, and confirmed the roles of the central coordinating committee, the working sub-committees on corpora and labeling, speech recognition assessment, and speech synthesis assessment, and continuation of the efforts in the future. Less