Question answering in Natural Language Processing for business applications
Reference persons GIUSEPPE CARLO CALAFIORE
Research Groups SYSTEMS AND DATA SCIENCE - SDS
Thesis type APPLIED RESEARCH
Description Question answering is a downstream application for natural language processing models. QA consists in processing large quantity of reference text data (in the form of a number of different documents) and a specific question with the appropriate NLP tools in order to find the answer to the question using the reference text data. QA models cover a number of different types of questions (yes/no questions, undefined form answers …) leading to a number of different type of answers (extractive answers, exact answers, abstractive answers etc).
In this thesis we are interested in developing and implementing efficient question answering tools focused on providing “first pass” answers to well defined questions, on the basis of an eterogeneous set of documents.
Required skills Informatics, machine learning, databases
Deadline 02/10/2021 PROPONI LA TUA CANDIDATURA