Design of Machine Translator based on QNN

Abstract

Machine translation (MT) along with natural language processing (NLP) always remained an area of interest for researchers since the computers were invented. Many researchers have tried to build the system which can understand multiple languages to translate from one source language to another target language. They also searched the way how computer understand and generate the human languages with semantics and syntactic. However, they realized that still many languages have translation difficulties, grammatically and semantically. Machine translation is a field of natural language processing. It involves the complete linguistic analysis of sentence used for automatic translation from one language to another. The main challenging issues need to be addressed are word ambiguity, word order, word sense, idioms, pronoun resolution, syntactic ambiguity and structural ambiguity. Recently some work has been done with Hindi to English and vice versa by several researchers using different methods of machine translation, like example based system, rule based, statistical machine translation, and parallel machine translation system. Some researchers have described the use of corpus pattern for alignment and reordering of words for English to Hindi machine translation using the neural network, but still there are a lot of possibilities to develop a MT System for Hindi to increase the accuracy of MT. This work presents the machine learning based translation system for Hindi to English and vice versa, which learns the semantically correct corpus. The quantum neural based pattern recognizer is used to recognize and learn the pattern of corpus using the information of parts of speech of individual word in the corpus like a human. The system performs the machine translation using its knowledge gained during the learning by inputting the pair of sentences of Devnagri (Hindi) and English. To analyze the effectiveness of the proposed approach, 2600 sentences have been evaluated during simulation and evaluation. The accuracy of proposed quantum neural network based machine translation system for Devanagari (Hindi) to English has been compared on different scores viz. BLEU, NIST, ROUGE-L, METEOR and human based evaluation, the accuracy are respectively 0.7502 on scale of 1, 6.5773 on scale of 10, 0.9233 on scale of 1, 0.5460 on scale of 1 respectively and 98.154 %. In case of English to Hindi MT system the accuracy achieved on BLEU, NIST, ROUGE-L, METEOR and human based evaluation respectively are 0.9809 on scale of 1, 7.3066 on scale of 10, 0.9887 on scale of 1, 0.9655 on scale of 1 and 98.261%. The accuracy of proposed system for both Hindi to English and English to Devanagari (Hindi) are found to be significantly higher in comparison with the existing English to Devanagari (Hindi) and Devanagari (Hindi) to English MT system like Google and Bing, Artificial Neural Network (ANN) based MT system and Anuvadaksh. The proposed system also learns and recognizes the Parts of Speech (POS) Tagging Pattern of English and Hindi corpora using the Quantum Neural Network based pattern recognition. To analyze the effectiveness of the proposed approach, 2600 sentences of news items having 11500 words from various newspapers have been evaluated. During simulations and evaluation, the accuracy of 98.40% for English POS Tagger and 99.13% for Hindi POS tagger has been achieved, which is significantly better in comparison with other existing approaches for Hindi and English parts of speech tagging.