Introduction
- Dr. Hab (Grenoble University)
- Language is still a new frontier for revolutionary ideas. Experience research in a friendly atmosphere open to original opinions.
- WASEDA University Researchers Database
- Email : yves.lepage(at)waseda.jp
*Swap “(at)” with “@”
Research
Example-based machine translation/NLP
The web is at the origin of an explosion of information called infoplosion. In particular, the number of texts in different languages increases everyday. For companies to access new markets, internationalization and localization have become a need. For search engines or survey agencies, as well as for the final user or the e-citizen, access to information in foreign languages has become a necessity. Natural language processing has developed knowledge-based and statistical methods. Knowledge-based methods take into account the productivity of language, but are uneasy to apply. Statistical methods are easy to apply, but fail at dealing with the productivity of language. Example-based methods try to take the best of two worlds: ease of application and account for productivity. In our research we want to formalize some example-based techniques, like analogy, and apply them to natural language applications like: automatic production of dictionaries, alignment, paraphrasing, machine translation.
- Machine Translation
- Natural Language Processing
- Learning by Examples, Example-based Methods
- Differences and Similarity, Analogy