NSOAMT

New Search Only Approach to Machine Translation

Project Page
Translation statistics

Our Goals

The aim of this project is to overcome the slowness and inaccuracy existing in translation semantics, through the development of a solution, which, based on the indexing of an incremental set of words, which will combine a certain semantic meaning, makes it possible to create a correspondence process, between its registration in the native language and the language of translation, so that the translation activity results in a relatively instantaneous and accurate process in the translation semantics.

Research Principle

This research principle is based on the assumption that the vocabulary used in a given type of work/document (ex: technical documentation) is relatively limited in terms of language style and word diversity, which enhances the greater speed and rigor of the translation process, through the developed/investigated indexing process.

Contribution to Society
The contribution of the new solution will bring about changes in the day-to-day life of society, assuming importance in current collective and individual life. The advantage of its dissemination will contribute to simplifying processes (whether administrative, legal, etc) and reducing associated costs, also contributing to streamlining the relationship between citizens and companies. In this way, it is expected to contribute to the proper functioning of different sectors (information services; governments; large multinationals; NGOs; etc), in order to reduce translation preparation time without sacrificing quality, allowing:

  1. Automate translations of repeated phrases (whether in the text itself or in previous ones);
  2. Suggest translations of similar phrases;
  3. Allow immediate search in the translation memory;
  4. Allows you to create terminology files integrated with the translation;
  5. Allow managing a translation memory with external resources;
  6. Facilitate the placement of the translation text in the final file;
  7. Increase the IT security of the translated text.

Published Results

September 2023, NSOAMT – New Search Only Approach to Machine Translation, arXiv:2309.10526

Co-Sponsors