Results - Details
Search command : Author="Φλουρής"
And Author="Γεώργιος"
Current Record: 8 of 22
|
Identifier |
000421549 |
Title |
Open domain question answering over hundreds of linked open datasets |
Alternative Title |
Aπάντηση ερωτήσεων ανοιχτού πεδίου πάνω από εκατοντάδες ανοιχτά συνδεδεμένα σύνολα δεδομένων |
Author
|
Δημητράκης, Ελευθέριος Θ.
|
Thesis advisor
|
Τζίτζικας, Γιάννης
|
Reviewer
|
Πλεξουσάκης, Δημήτρης
Φλουρής, Γεώργιος
|
Abstract |
Open domain Question Answering is a challenging task that requires, among others, to tackle the
data distribution issue, i.e. the fact that datasets are scattered in several places. In this thesis, we
focus on open domain Question Answering over Linked Open Data. We confine ourselves to three
kinds of questions: factoid, confirmation, and definition questions. We introduce and
comparatively evaluate information extraction based processes for question answering. The
distinctive feature of our approach is that it can answer questions over millions of entities, by
exploiting hundreds of Linked Data sources simultaneously, without having to use any training
data. The process comprises three main phases: (i) Question Analysis (that includes question type identification and cleaning), (ii) Entities
Detection, Named Entity recognition and linking and (iii) Answer Extraction (that includes RDF
triples retrieval, scoring and matching). We demonstrate the benefits of this approach in terms of
answerable questions and answer verification, and we investigate, through experimental results,
how the steps of the question answering process affect the effectiveness of question answering.
The evaluation was based on 1000 questions from SimpleQuestions and 2500 from QALD-7 Largescale datasets.
|
Language |
English |
Subject |
Factoid questions |
|
Open domain |
|
Question answering |
|
Ανοιχτό πεδίο |
|
Απάντηση ερωτήσεων |
|
Ερωτήσεις γεγονότων |
Issue date |
2019-03-29 |
Collection
|
School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses
|
|
Type of Work--Post-graduate theses
|
Permanent Link |
https://elocus.lib.uoc.gr//dlib/5/e/7/metadata-dlib-1552037158-385984-27635.tkl
|
Views |
618 |