Question Answering Module Leveraging Heterogeneous Datasets

Question Answering has been a well-researched NLP area over recent years. It become necessary for users to be able query through the variety of information available - it structured or unstructured. In this paper, we propose module which a) can consume data formats heterogeneous pipeline, ingests from product manuals, technical forums, internal discussion groups, etc. b) addresses practical challenges faced in real-life situations by pointing exact segment manual chat threads solve user c) provides segments texts when deemed relevant, based on and business context. Our solution comprehensive detailed pipeline that is composed elaborate ingestion, parsing, indexing, querying modules. capable handling plethora sources such as text, images, tables, community flow charts. studies performed business-specific datasets represent necessity custom pipelines like proposed one several real-world document question-answering.