AUTOMATIC QUESTION ANSWERING IN MYANMAR LANGUAGE USING SEQUENCE TO SEQUENCE WITH ATTENTION MECHANISM
Naing Naing Khin, Khin Mar Soe · Indian Journal of Computer Science and Engineering · 2024
The primary and enduring objective of natural language processing and artificial intelligence is teaching machines to read and comprehend human natural language.Question Answering is a conversational model which can generate an answer automatically from a given context.In many systems, they have great potential for those interested in learning the necessary details.QA can be based on various techniques, including information retrieval, knowledge-based, generative, and rule-based approaches.Sequence to Sequence (Seq2Seq) model is adapted for good conversational model for QA.There are many Question Answering Models in English but there is still need a system that can provide convenient question and answer in Myanmar Language.In this research, we proposed Automatic Question Answering in Myanmar language utilizing the RNN encoder-decoder model-based Sequence to Sequence model with Attention Mechanism.34,850 Myanmar language question-answer pairs are used to train the question answering model.The experimental result is evaluated on Precision, Recall, F1 Score, Human Evaluation and gets overall accuracy 71%.