PhD Thesis
When the Answer comes into Question in Question-Answering
2013
—Key information
Authors:
Supervisors:
Published in
10/16/2013
Abstract
A Question-Answering system aims at returning correct answers to questions posed in natural language. The typical architecture of a Question-Answering system includes a component dedicated to Answer Extraction, which is the focus of this thesis. This thesis starts by presenting Just.Ask, the Question-Answering system developed during the course of these Ph.D. studies, that represents the baseline and the target of all the experiments. Afterwards, the tasks of candidate answer extraction and final answer selection are addressed. The former uses (lexico-syntactic) patterns, automatically learned from largescale information sources using a minimally supervised approach. In the latter, semantic relations – equivalence and inclusion – are detected between the candidate answers and used to better choose the final answer. Finally, a new approach to the Question-Answering challenge is presented, making Just. Ask the first system, to the best of our knowledge, which uses the correct answers to past questions to answer future questions, employing the learned patterns. In this approach, the user has a fundamental role that allows the learning process to be triggered: s/he confirms the correctness of the system’s answer(s). However, if none of the returned answers is correct, the user can provide the correct answer to the posed question.
Publication details
Authors in the community:
Ana Cristina Bastos Mendes
ist150951
Supervisors of this institution:
RENATES TID
101296584
Degree Name
Doutoramento em Engenharia Informática e de Computadores
Fields of Science and Technology (FOS)
electrical-engineering-electronic-engineering-information-engineering - Electrical engineering, electronic engineering, information engineering
Keywords
- Question-Answering Systems
- Candidate Answer Extraction
- Lexico-syntactic Patterns
- Pattern Learning
- Pattern/Sentence Unification
- Relaxation Strategies
- Final Answer Selection
- Semantic
- Relations
- Learning to Answer
- User Feedback
Publication language (ISO code)
eng - English
Rights type:
Embargo lifted
Date available:
08/11/2014
Institution name
Instituto Superior Técnico