Profile
Fig. 1 - Hideo Watanabe
Hideo Watanabe is a manager of Knowledge Infrastructure group of Tokyo Research Laboratory, IBM Japan.
He joined Tokyo Research Laboratory in 1986, and has been working on natural language processing research, in particular on machine translation. In addition to research work, He contributed on the development of IBM machine translation products; Internet King of Translation and WebSphere Translation Server.
He received Ph.D of engineering from Kyoto University in 1996. He is a member of Information Processing Society of Japan, Association of Natural Language Processing, and Japanese Society for Artifical Intelligence.
Contact:
e-mail: 
Research Topics
He has been working on machine translation research. In particular, his main contributions are exmaple-based approaches for machine translation. He has developed an example-based transfer system called "SimTran" which finds similar translation examples for an input sentence, and contructs an output sentence by combining target language information of selected similar translation examples.
He has also studied about an approach for automatic text summarization, which is basically a type of summarization mechanism for picking up important sentences, but the importance calculation policy is changed based on article types or domains.
The recent research topic of his group is text-mining. A text-mining system called IBM TAKMI has been studied in Knowledge Infrastructure group since 1997, and it has been productized as OmniFind Analytics Edition in 2007 from IBM. (Now it is renamed as IBM Content Analyzer.) This group has been continuing text-mining research and is expanding the scope of the research from contact center log to Internet and Conversation data.