1. Knowledge Enabled Information and Services Science Ontology supported Knowledge Discovery in the field of Human Performance and Cognition Kno.e.sis Center http://knoesis.org Wright State University C. Thomas, P. Mendes, D. Cameron, A. Sheth, TK Prasad Thanks, C. Ramakrishnan

2. Knowledge Enabled Information and Services Science Motivation Speed up the “search” part of research Develop a system that helps in focused exploration of a domain

3. Knowledge Enabled Information and Services Science Cognition-related concepts

4. Knowledge Enabled Information and Services Science Subject  Predicate  Object search

5. Knowledge Enabled Information and Services Science Subject  Predicate  ? search

6. Knowledge Enabled Information and Services Science Demo

7. Knowledge Enabled Information and Services Science Motivation

8. Knowledge Enabled Information and Services Science Motivation For this reason we propose a largely automatic domain model generation on the basis of existing knowledge repositories and text corpora, namely MeSH, MedLine and Wikipedia.

9. Knowledge Enabled Information and Services Science Results

10. Knowledge Enabled Information and Services Science Full Class Hierarchy

11. Knowledge Enabled Information and Services Science Top Levels by importance (#Concepts)

12. Knowledge Enabled Information and Services Science Selective Hierarchy – CogSci/NeuroSci

13. Knowledge Enabled Information and Services Science Zooming in on Cognition-related concepts

14. Knowledge Enabled Information and Services Science Biomolecules-related concepts

15. Knowledge Enabled Information and Services Science Technologies

16. Knowledge Enabled Information and Services Science Steps 1.Carve a focused domain model out of Wikipedia 2.Identify mentions of entities and relationships in the relevant scientific literature (Pubmed abstracts) to support non-hierarchical guidance. 3.Map extracted entity mentions to concepts and extracted predicates to relationships 4.Applications to access research literature guided by the domain model.

17. Knowledge Enabled Information and Services Science Steps 1.Carve a focused domain model out of Wikipedia 2.Identify mentions of entities and relationships in the relevant scientific literature (Pubmed abstracts) to support non-hierarchical guidance. 3.Map extracted entity mentions to concepts and extracted predicates to relationships 4.Applications to access research literature guided by the domain model.

18. Knowledge Enabled Information and Services Science

19. Wisdom of the Crowds

20. Knowledge Enabled Information and Services Science

21. Overview - conceptual

22. Knowledge Enabled Information and Services Science Expand - conceptually Full text search on Article texts Delete results with low Lucene score Graph-based expansion

23. Knowledge Enabled Information and Services Science Node Similarity computation

24. Knowledge Enabled Information and Services Science Collecting Instances

25. Knowledge Enabled Information and Services Science Creating a Hierarchy

26. Knowledge Enabled Information and Services Science Creating a Hierarchy

27. Knowledge Enabled Information and Services Science Hierarchy Creation - summary

28. Knowledge Enabled Information and Services Science Reduce steps Remove all terms that have low pertinence to the domain Intersect hierarchy with broader focus domain Reduce hierarchy depth

29. Knowledge Enabled Information and Services Science Remove unwanted individuals

30. Knowledge Enabled Information and Services Science Remove unwanted categories

31. Knowledge Enabled Information and Services Science Flatten categories

32. Knowledge Enabled Information and Services Science Snapshot of final Topic Hierarchy

33. Knowledge Enabled Information and Services Science Evaluation wrt. expert model Evaluation wrt. MeSH versions of 2004 (04) and 2008 (08) for both the restricted and the full set of MeSH term

34. Knowledge Enabled Information and Services Science Steps 1.Carve a focused domain model out of Wikipedia 2.Identify mentions of entities and relationships in the relevant scientific literature (Pubmed abstracts) to support non-hierarchical guidance. 3.Map extracted entity mentions to concepts and extracted predicates to relationships 4.Applications to access research literature guided by the domain model.

35. Knowledge Enabled Information and Services Science Challenges and Opportunities Vocabularies, Thesauri, Ontologies, exist in several fields –How can we use them? lexicon: Fish Oils, Raynaud’s Disease, etc. types/labels: Fish Oils instance of Lipid relationships between types: Lipid affects Disease Identification of simple ontology terms in text is not enough –Compound Entities –Complex Relationships 35 But

36. Knowledge Enabled Information and Services Science Challenge: Compound Entities 36 Entities (MeSH terms) in sentences occur in modified forms “ adenomatous ” modifies “ hyperplasia ” “ An excessive endogenous or exogenous stimulation ” modifies “ estrogen ” Entities can also occur as composites of 2 or more other entities “ adenomatous hyperplasia ” and “ endometrium ” occur as “ adenomatous hyperplasia of the endometrium ” MeSH termsModifier

37. Knowledge Enabled Information and Services Science 37 Extraction Algorithm Relationship head Subject head Object head

38. Knowledge Enabled Information and Services Science Extraction Approach ● Parse sentences with a dependency parser ● Use a few domain-independent rules to segment sentences into Subj  Pred  Object ● Subjects and Objects represent compound entities

39. Knowledge Enabled Information and Services Science 39 Preliminary results Subject types, inferred from the HEAD of the compound

40. Knowledge Enabled Information and Services Science 40 Extracted Triples

41. Knowledge Enabled Information and Services Science 41 Representation – Resulting RDF Modifiers Modified entities Composite Entities

42. Knowledge Enabled Information and Services Science Using the generated RDF

43. Knowledge Enabled Information and Services Science Steps 1.Carve a focused domain model out of Wikipedia 2.Identify mentions of entities and relationships in the research literature. 3.Map extracted entity mentions to concepts and extracted predicates to relationships 4.Applications to access research literature guided by the domain model.

44. Knowledge Enabled Information and Services Science Step 3: Identifying Synonymy and Antonymy between Verbs for Relationship Matching Christopher Thomas, Wenbo Wang

45. Knowledge Enabled Information and Services Science Motivation An ontology has clearly demarcated, formal relationships whereas natural language uses multiple ways of representing these relationships Focus on verbs expressing relationships NLP-extracted triples use verbs to indicate relationships, but are not aligned with the relationships yet. Align similar verbs  Align extracted predicates to formal relationships

46. Knowledge Enabled Information and Services Science Examples of relationships to merge Dilation of the fourth ventricle indicated atrophy of the middle cerebellar peduncle. –E.g.  indicate The lesions in the rolipram-treated group also showed increased astrogliosis and increased CREB phosphorylation in the activated microglia and astrocytes. –E.g.  showed GOOD BAD  indicate GOOD

47. Knowledge Enabled Information and Services Science Short Review – training set creation Normalization POS Tagging Synonyms & antonyms extraction Pattern extraction Pattern2Relationship matrix VerbPair2Pattern matrix Remove phrases containing less than 2 verbs Eating ->, eats -> ANT: Like Hate SYN: calculate compute He learns and memorizes materials What is the correlation of this pattern and that relationship? What is the correlation of this verb pair and that pattern?

48. Knowledge Enabled Information and Services Science Design Synonyms & antonyms extraction from Wordnet –A single verb has multiple meanings contract, take, get (be stricken by an illness, fall victim to an illness) "He got AIDS"; "She came down with pneumonia"; "She took a chill“ film#1, shoot#4, take#16 (make a film or photograph of something) "take a scene"; "shoot a movie" Take has 42 meanings in Wordnet –Some meanings of a verb are less frequently used –If the number of meanings of a verb exceeds threshold, we will abandon this verb –If the frequency of a specific meaning of a verb is low, we will abandon this meaning

49. Knowledge Enabled Information and Services Science Solution Do pattern-based text analysis to find predicates that are used in similar contexts and map them to upper-level relationships, e.g. those found in UMLS

50. Knowledge Enabled Information and Services Science Design Pattern Extraction (when meet synonyms and antonyms) –List all the possibilities –He learns and memorizes materials and and materials He and He and materials –Each pattern has three counts: Syn, Ant, Other and Displays and exhibits -> increase Syn count Likes and hates -> increase Ant count Dances and sings -> increase Other count

51. Knowledge Enabled Information and Services Science Design Pattern2Relation Matrix(3 * 180363) Observation –Count number for Other Relation is greater than counts of Syn and Syn in most times. Do row based normalization (Smoothing) And then do column based normalization [p(Rel|Pattern)] –Some patterns occurs a few times in matrix Column based minimization and either or neither nor Syn469201 Ant15177501188 Other8408391553519

52. Knowledge Enabled Information and Services Science Design VerbPair2Pattern Matrix –Count the number of patterns occurring with every verb pair Very sparse(many zeroes in matrix) Size:887781 * 180363 Some verb pairs occur with a few patterns(not statistically significant) Row based Minimization and either or neither nor hate like61214 struggle fight123 suffer enjoy000

53. Knowledge Enabled Information and Services Science Design VerbPair2Pattern Matrix –Matrix verbpair2pattern * Matrix pattern2relation SynAntOther struggle fight0.3079031370.5534522030.13864466 free release0.56049579700.439504203 permit forbid0.2292655820.5750810390.195653379

54. Knowledge Enabled Information and Services Science Corpus Statistics

55. Knowledge Enabled Information and Services Science Corpus Statistics Zipf- Distribution of Verb Pairs

56. Knowledge Enabled Information and Services Science Evaluation Obtain relationships of patterns and semantic relations, then apply them to Wikipedia Criteria(take SYN for example) –Strict(SYN con > 0.5) –Moderate (SYN con > ANT con AND SYN con > OTHER con ) –Loose(SYN con > ANT con )

57. Knowledge Enabled Information and Services Science Evaluation

58. Knowledge Enabled Information and Services Science In Progress Once the evaluations show constant results, apply Synonym finding to the predicates extracted by the NLP-based relationship extraction

59. Knowledge Enabled Information and Services Science Steps 1.Carve a focused domain model out of Wikipedia 2.Identify mentions of entities and relationships in the research literature. 3.Map extracted entity mentions to concepts and extracted predicates to relationships 4.Applications to access research literature guided by the domain model.

60. Trellis: Knowledge-Driven Web Browsing Pablo N. Mendes, Cartic Ramakrishnan, Delroy Cameron and Amit P. Sheth Knoesis Center, May 27 2009. Step 4

61. Introduction  Background  System Overview  Demo  Future Work 61

62. Background  Web Browsing  Search  Keyword search  Point of access to related content  Sift  Retrieve documents  Hyperlinks  Query Refinement Search Sift 62 and

63. Background  Observations 63 Human Cognition Model Information Need (Documents) Keyword Entity e1e2 Relationship e1e1 enen Pathway Graduate Student Research Assistant Turkey San Diego

64. System Overview  New Browsing Paradigm  Search  Navigate Context  Organize results  Trellis  Intricate support structure for vines  Vines as information need/context through navigation 64 Search Navigate/ Stitch Share

65. System Overview 65 Knowledge Base {Pubmed Abstracts} Knowledge Base {Pubmed Abstracts} Spotter Index Workbench Save/Publish Ontology {NLM hand annotated triples} Ontology {NLM hand annotated triples} Fig.1 Trellis Architecture

66. Magnesium -> isa -> Calcium Channel Blocker

67. Calcium Channel Blockers -> prevent -> Migraine

68. Magnesium -> isa -> Calcium Channel Blocker Migraine -> prevent -> Migraine Magnesium -> prevent -> Migraine

69. Future Work  Ranking - Boosting Precision/Recall  User feedback  Corpus Statistics  Scalability  Information Extraction from natural language  Named Entity Recognition  User driven  Knowledge Discovery 69

70. Knowledge Enabled Information and Services Science Thank you