1. Question Answering Based on Semantic Graphs Lorand Dali – lorand.dali@ijs.si Delia Rusu – delia.rusu@ijs.si Blaž Fortuna – blaz.fortuna@ijs.si Dunja Mladenić – dunja.mladenic@ijs.si Marko Grobelnik – marko.grobelnik@ijs.si

2. Motivation System Overview Question Answering Document Overview Facts Semantic Graph Document Summary Conclusions Overview

3. Motivation

5. Triplets Facts stated in the text The core of the sentence (subject, verb, object)

6. System Overview

7. Extract facts (triplets) from text Index triplets to enable structured search on them Analyze questions to obtain the queries for the triplet search Retrieve the answer and the document containing it Browse the document overview Question Answering

10. Question types: Yes/No questions (Do animals eat fruit?), list questions (What do animals eat?), reason questions (Why do animals eat fruit?), quantity questions (How much fruit do animals eat?), location questions (Where do animals eat?) and time questions (When do animals eat?). Question Answering

11. Analyze the document containing the answer: Highlight facts described by subject – verb – object triplets (identified in the Penn Treebank parse tree) Obtain the document semantic graph View the automatic document summary Document Overview

12. Semantic Graph Document Plain text format Named entity extraction Co-reference resolution According to traditional Chinese medical belief, mental problems, laziness, malaria, epilepsy, toothache and lack of sexual appetite can be treated with tiger parts, leading to rampant poaching of the animal in Asia, the World Wide Fund ( WWF ) said. Asia World Wide FundWWF Asia - location World Wide Fund - organization WWF -organization Co-reference S – V – O triplet extraction Triplet enhancement Semantic Graph

13. Document Summary Feature Extractor Features: linguistic document graph Linear SVM Linear Model The Kerinci conservation project, an area of around three million hectares (7. 4 million acres) in west Sumatra, was being supported by funds from the World Bank, Subijanto said. [10.0912] Subijanto, a spokesman for the Forestry Ministry, said Indonesia was commited to protecting the tigers, which live within Sumatra's four designated conservation areas. [9.4155] Ranking

14. Document Summary There are people wanting tiger products who didn't want them before, " Ron Lilley,coordinator for species conservation at the WWF in Jakarta, told Reuters. Subijanto, a spokesman for the Forestry Ministry, said Indonesia was commited to protecting the tigers, which live within Sumatra's four designated conservation areas. The Kerinci conservation project, an area of around three million hectares (7. 4 million acres) in west Sumatra, was being supported by funds from the World Bank, Subijanto said.

15. Enhanced question answering system Question answering, where the answer is supported by documents Document browsing Facts Document semantic graph Automatic document summary Conclusions

16. Future work System extensions: triplet extraction, named entity recognition Expand the search to look for answers in ontologies Relax the requirement that the questions have a predefined form Improve the document overview functionality by integrating external resources Conclusions

17. Thank you! Questions are guaranteed in life, answers aren’t.

18. Extracted features: Document Summary Linguistic Attributes (13)Document Attributes (11) Graph Attributes (9) Logical form tag Treebank tag Part of speech tag Depth of linguistic node 8 semantic tags for named entities Sentence related: e.g. – location of sentence within doc Triplet related: e.g.- frequency of triplet element in sentence, in doc, … Authority and Hub weight, Page Rank Node degree Size of weakly connected component Size of max length chain Frequency of verbs among edges

19. Document Summary Object - WordSubject - WordVerb - WordLocation Of Sentence In DocumentSimilarity With CentroidNumber Of Locations In SentenceNumber Of Named Entities In SentenceAuthority Weight ObjectHub Weight SubjectSize Weakly Conn Comp Object Rank (Information Gain)