Applications

Morpho-syntactic and semantic analysis of texts, like the ones presented in the previous chapters, are usually performed in order to prepare the text for an precise application. The goal of the current chapter is focus on these applications of NLP.

Just like in the USA or in Europe, the main efforts in Japan are focused on applications that will enable the computers to have a high added value compared with human in the field of textual document manipulation and processing. Among these applications with huge potential economical rewards we can quote the following:

• information retrieval and extraction in a large textual database;
• automatic classification of documents;
• alignment of several information sources (e.g. newspapers, television, radio...);
• disambiguation of speech or handwriting recognition;
• automatic summarization of texts;
• machine translation (MT);
• human/machine interfaces.

• Information retrieval and extraction
  • Request expansion at Tokodai
  • Geographical knowledge organization at NTT
  • 5W1H classification at NEC
  • Information extraction with template matching at Kyutech
  • Use of a concept base at NTT
• Document classification
  • Kanjis retrieval at Kyoto University
  • Automatic document classification at Tokushima University
• Multimedia alignment at Kyoto University
• Characters disambiguation
  • N-gram models at Kyoto University
  • A mixture of n-gram models at Tohoku University
• Automatic Summarization
  • The system TESS at Kyutech
• Machine translation (MT)
  • NEC's PIVOT
  • NTT's ALT-JE
  • ATR's MATRIX
• Human/machine interface
  • A human/machine dialogue system at NTT
  • The competition DIALEAGUE organized by ETL
  • IRENA system at Kyutech