Quantum neural network based machine translator for Hindi to English

Area of Science:

• Natural Language Processing
• Artificial Intelligence
• Computational Linguistics

Background:

• Traditional machine translation systems often struggle with semantic nuances in Hindi to English translation.
• Existing systems may lack the ability to learn and adapt to the complexities of linguistic patterns.

Purpose of the Study:

• To develop and evaluate a novel machine learning-based machine translation system for Hindi to English.
• To leverage quantum neural networks for enhanced pattern recognition and semantic understanding in translation.

Main Methods:

• The system employs a quantum neural-based pattern recognizer to learn from a semantically correct corpus.
• It incorporates part-of-speech information for individual words, mimicking human linguistic processing.
• Machine translation is performed by inputting sentence pairs in Devnagri-Hindi and English.

Main Results:

• The proposed system achieved high scores on standard translation evaluation metrics: BLEU (0.7502), NIST (6.5773), ROUGE-L (0.9233), and METEOR (0.5456).
• Performance significantly surpassed that of Google Translation and Bing Translation for Hindi to English.
• Evaluation was conducted on a dataset of 2600 sentences.

Conclusions:

• The quantum neural network approach offers a significant improvement in Hindi to English machine translation.
• The system demonstrates effective semantic learning and pattern recognition capabilities.
• This method presents a promising advancement over current state-of-the-art translation services.