We build a QA (Question Answering) system based on Deep Learning that understands a question and deliver a single correct answer from a large unstructured data-set. Our system automatically answer questions posed by human in natural language query format. By year 2020, 50XB of data will be created mainly due to the increase of unstructured data. One of our deep learning engine called MRC (Machine Reading Comprehension) can handle an unstructured passage to find answers. As a public reference with the accuracy of our MRC engine, we have co-ranked #1 with Google AI team in Stanford University MRC competition called SQuAD 2.0 (as of Nov. 2018).
Our QA system targets industries such as 1)Data Intelligence for business service, 2)AI-based chatbots and 3)Smart devices by leveraging deep learning engines such as NLU(Natural Language Understanding), MRC(Machine Reading Comprehension) and Paraphrasing engines to help optimize workflows with data-intensive companies. MRC (Machine Reading Comprehension) is an AI technology that understands a user’s questions based on deep learning and provides a single answer. It finds the correct answer from unstructured data instantly, thus, there is no need to build an ontology or data formatting. As a public reference with our MRC engine we have ranked no. 1 with Google AI team in SQuAD 2.0 competition by Stanford University’s NLP team.