My Master's Dissertation
This dissertation aims to handle the task of Community Question Answering (CQA) through learning contextual representation of data by implementing deep neural models. Primarily, sequence modeling is experimented for effective deduction or selection or classification of “relevant” responses for a posed question. In CQA, open-domain user interaction creates abundance of dynamic data that is vital to the forum. Such websites require efficient, fast and accurate knowledge which is achieved by finding the most relevant answers for them and so forth, exhibiting them by their order of relevance to the users. The work done in this dissertation achieves the same using Natural Language Processing and Deep Learning techniques. As sequential networks have been helpful for understanding both the textual features and hidden context of the text, they are applied to acquire meaning from the forum’s data.
Keywords: Community Question Answering (CQA), Open-domain Question Answering (QA), Natural Language Processing (NLP), Sequential networks, Long Short-term Memory, Word Embeddings, Answer Selection
Feel free to reach me for details of experiments.