(N)MT Project: A Case Study

Within its historical development, there have been attempts to create the ideal machine translation engines and to improve it using various methods, and the “statistical and rules-based” experiments have stood out in the process.

Lately, the machine translation has become a much more popular issue in the world as the codes for the statistical machine translation engine, Moses, have been opened to the public, and Google Inc. has started offering its machine translation service free of charge.

However, failing to provide adequate performance on languages with very rich morphology, such as Turkish, the machine translation made a sensation in our country with Neural Machine Translation (NMT) method, which is trying to imitate the operating system of human brain.

In this presentation, the team that developed the first and only Neural Machine Translation (NMT) engine in Turkey and become one of the firsts in the world will relay the emerging developments, how they could change the translation industry and their own practical usage.