Understanding the Subjective Aspect of Question Answer

Patel, Ashish (2020) Understanding the Subjective Aspect of Question Answer. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (846kB) | Preview
Preview	PDF (Configuration manual) Download (1MB) | Preview

Abstract

Question answer system these days are good when it comes to fact based or verifiable answers but when it comes to questions which seek recommendations, personal experiences or opinions, humans are much better at answering. One could say that humans are great at solving contextual problems that need a broader, multidimensional view of the context, something machines are not qualified to do well yet, as questions could be of several forms like multi sentence elaborations while others could be incomplete without proper context. Unfortunately, it is very hard to build a good subjective question answer system due of the lack of trained data. To rectify the problem, Google with the help of it’s crowdsource team came out with a dataset which comprises of question answers pairs from various open source websites which are given scores between 0 to 1 on 30 different subjective aspect of the question answer pair like question is well written or not, answer provided is satisfactory or not etc., rated by the team itself. The aim of this research is to take the above dataset and create a model which could be able to score the question answer pair of their subjective aspect. To achieve the above results three different NLP techniques were used Word Embeddings, Universal Sentence Encoder and BERT transformer model and their results were compared. Throughout the result it was found that against the BERT model, which is considered gold standard in NLP, Universal sentence encoder gave equal if not better result for the data set.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Computer software T Technology > T Technology (General) > Information Technology > Computer software
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Dan English
Date Deposited:	25 Jan 2021 13:44
Last Modified:	25 Jan 2021 13:44
URI:	https://norma.ncirl.ie/id/eprint/4461

Actions (login required)

View Item