Published Paper Details:

Language, being the carrier of information, holds utmost importance for human communication and information sharing. Difficulties arise when individuals from different language backgrounds try to interact, leading to language barriers. To address this issue, the need for efficient and rapid translators becomes evident. At first, human translators played this function, but they are not a lasting answer. Consequently, scholars have carried out numerous tests and attempts, leading to the emergence of machine translation technology. This technology has been widely utilized, exploring diverse languages, and employing various methodologies like Rule-Based, Statistical-Based, and more recently, Neural Machine Translation. Machine Translation has been a familiar notion for the last five decades, falling within the domain of Natural Language Processing (NLP), a subfield of Machine Learning. The effectiveness of machine translation heavily relies on the quality of the datasets and the training the machine has undergone. However, many challenges have emerged due to the lack of adequate datasets for certain languages, often referred to as low-resource languages. Our project primarily focuses on translating sentences, poems, and proverbs from South East Indian Languages, specifically Telugu, into English. As previously mentioned, Telugu is considered a low-resource language and is notoriously challenging for human learners. It boasts a wide array of proverbs and poems, but unlike English and other languages, Telugu has not been extensively translated. Previous attempts to translate Telugu poems and proverbs have yielded unsatisfactory results. Our main objective, therefore, is to develop a machine translation model using OpenNMT that can accurately translate Telugu sentences, poems, and proverbs into English.