Gittens Web site
Font size:      

The Mimida Project Home Page

The logo of the Mimida project; Thanks Michelle!

Mimida; A mechanically generated Multilingual Semantic Network

You are in the right place for information about the Mimida multilingual semantic network. This project is part of an exercise in knowledge representation and an exponent of my interest in language. The Mimida project is one the projects that compete for my spare time.

Languages included in Mimia

The Mimida Multilingual Semantic Network is a mechanically generated Semantic Network based on the WordNet Semantic Network and a number of dictionaries I found on the Web. It currently includes information words from the following languages.

  • afrikaans
  • albanian
  • catalan
  • danish
  • dutch
  • english
  • estonian
  • finish
  • french
  • fries
  • german
  • icelandic
  • irish
  • italian
  • latin
  • mayan
  • norwegian
  • oldenglish
  • papiamento
  • portugese
  • scottish
  • spanish
  • sranan
  • swahili
  • swedish
  • tahitian
  • tswana
  • zulu

The languages currently included in Mimida where arbitrarily selected and since I don't speak most of these languages I don't have a clue what the quality of the generated network is for most of these languages. The process I used to generate the Mimida multilingual Semantic Network is pretty straight-forward. I hope to describe the process in a paper soon.

Feel free to browse the current version of the Mimida Multilingual Semantic Network.

Room for improvement

The current version of the Mimida Multilingual Semantic Network clearly has room for improvement. Obvious points to be worked on:

  • make the knowledge representation system understand Unicode strings because currently a lot of input gets rejected because it is not ASCII (Partially done).
  • present different views of the information in the Multilingual Semantic Network
  • make the browser a bit more sexy and user friendly
  • add search interfaces for languages other than english; so that you can enter search strings in other languages as well
  • improve heuristics for matching synsets to include language specific knowledge
  • add plausibility tests for matches
  • fix bugs...
  • etc; etc.

The Mimida project will probably develop in a seriously easy-going fashion :-).

Please note that except for being a grateful user of the WordNet semantic Network, neither I or the Mimida project is in anyway related to the WordNet project or Princeton University.
This page is written and maintained by Maurice Gittens, feel free to check out my home page or to send me your comments.