Farahnaz Bagheri - Birjand University

Fully Automatic Machine Translation which means translating a wide variety of texts without human’s intervention was a dream in distant vision for human beings from one language into another language but, it came into reality and even achievable for scientific and technical documents through computer programs. Neural machine translation approach, as a promising approach departs phrase-based statistical MT but, is relatively new approach to statistical one based on neural networks. The researchers are less inclined to test it in MT system. The researcher in this study will be of interest to apply Neuro-Genetic, combination of Neural Network and Evolutionary Computation, to improve software’s ability in processing data through Genetic Algorithm to translate efficiently. Hence, this study concerns to introduce Soft Computing to optimize the quality in translation, whether to distinguish or to substitute the suitable choice, and lends the MT system a hand to tackle mistranslations of the constituent which makes output unnatural and lack of integrity.

Machine Translation, Neural Network, Genetic Algorithm