| YAN Yonghong
Professor, Supervisor of Doctor Candidates, Director of ThinkIT Lab and Director Assistant
Tel: (86-10) 62621927 Ext 1888
Yan Yonghong, graduated from the Electronic Engineering Department of Qinghua University in July 1990, and got his PhD in Computer Science and Engineering from OGI (Oregon Graduate Institute of Science & Engineering) in 1995. From 1995, he taught masters and doctoral students in the Oregon Research Institute , served as Associate Director in the Center of OGI For Spoken Language Understanding. He served as Director of Intel microprocessors laboratory division engineers and computer interface general framework, Intel China Research Center Director and Principal Engineer, Intel global computer interface academic Chairman of the Committee and Associate Professor of Oregon Health & Science University. In 2002 he was elected in "CAS scientists" and now a researcher at the Chinese Academy of Sciences Institute of acoustic, tutor of doctoral students, the Director and Assistant Director of the Laboratory ThinkIT. At the same time, he also served as the Chinese National Natural Science Fund Committee and the U.S. expert panel, the National "242" panel of experts, including a number of members post.
He engaged in human interface (especially language processing and voice recognition) research for many years and is the main author of OGI speech toolkit (CSLUSPEECHTOOLKIT) voice recognition part. In the voice system based on neural network framework (HMM and Neural net integration and training), modelling and training algorithms He innovated (using EM algorithm and training since the correlation coefficient of 2,463 non-0 or 1). He made many identifier end, multi-application and back-end information integration algorithm features in the areas of language identification greater impact study and peer widely used. In the U.S. Department of Defense and the Bureau of Standards in the annual competitions held in 1995 and 1996 , he earned them twicely . He led Intel integrated performance database (Intel Integrated Performance Primitives ( IPP s ) ), part of the development of voice. Since 1986 he has launched a major non-specific persons consecutive term table speech recognition research , making many innovatation in modelling, adaptive and search spheres, repeatedly made the first in the 863 voice recognition software. He has over 40 patents (including mandated and has been submitted to the Patent Office), 100 articles published papers. Since its inception in 2002, the laboratory has repeatedly won national organizations voice system software first.
Currently engaged in research areas:
large scale non-specific persons consecutive term voice recognition, voice signals end processing, multimedia data retrieval, audio watermarks, multimode language system, plug-in systems, adaptive systems and rapid search algorithm.