National and Kapodistrian University of Athens
Υπολογιστικό μεταγνωστικό σύστημα απάντησης ερωτήσεων εμπνευσμένο από μεταγνωσιακές λειτουργίες
Abstract
dc:descriptionThe objective of the thesis was the implementation of a Question Answering System with architecture based on the metacognitive management of strategies for answering questions and generating explanations justifying the answers. The system has a structure consisting of the following subsystems: 1. The subsystem of question processing that accepts questions expressed in Natural Language and reformulates them. 2. The subsystem of analysis of the text used for answering the questions that processes it and produces files with intermediate results. In one of these files the periods are stored separated in three parts: the part to the left of the verb, the verb, and the part to the right of the verb. 3. The subsystem of automatic extraction of knowledge which produces files that contain entities that are described by the text. 4. The subsystem of question answering that is tries the available strategies according to a concrete order of succession, and applies first that achieves the generation of an answer. The choice of a strategy concerns the reformulation of questions. The answer is either an explicit statement existing in the text or an implicit one that results from reasoning that combines prerequisite knowledge of the field and parts of the text. After the generation of an answer the user is able to watch the steps of reasoning that are justified by the text. The answers contain explanations that demonstrate the characteristic of self-awareness. 5. The subsystem of diagrams that offers the possibility to the user of inspecting diagrams that support the explanations of geometric proofs. Conclusions: The answering of questions is accomplished without translating them to any logical form. The knowledge used for answering the questions is in the form a natural language text. Explanations are produced that report the reasons of success or failure of the satisfaction of the objective of each stage of processing for answering the questions. The system can change the strategy of pursuing an objective when an answer is not achieved. Explanations are generated for strategy choice. The adaptation to different domains is possible with the change of Ontology and Dictionary only. The extension of the system is achieved simply by the addition of strategies and linguistic knowledge for the treatment of questions and text.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2010
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Αρμάος, Ιωσήφ
- Armaos, Joseph
Subjects
dc:subject × 20- Απάντηση ερώτησης από κείμενα
- Μηχανική συνείδηση
- Επεξήγηση
- Βιοϊατρικά κείμενα
- Απάντηση ερώτησης με συλλογισμό
- Συμπερασμός με κείμενα
- Συστήματα με αυτοεπίγνωση
- Γεωμετρική απόδειξη
- Question answering from texts
- Machine consciousness
- Explanation
- Biomedical texts
- Deductive question answering
- Inference with texts
- Self-aware systems
- Geometry proof
- Κοινωνικές Επιστήμες
- Άλλες Κοινωνικές Επιστήμες
- Social Sciences
- Other Social Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/28760