Ontologies Growing Up: Tools for Ontology Management. Natasha Noy Stanford University

Transcription

1 Ontologies Growing Up: Tools for Ontology Management Natasha Noy Stanford University 1

2 An ontology Conceptualization of a domain that is formal can be used for inference makes assumptions explicit shared, agreed upon enables knowledge reuse facilitates interoperation among applications and software agents 2

3 An ontology (II) Wine produced_by Winery subclass-of subclass-of Defines classes, properties, and constraints in a domain White wine subclass-of Red wine tannin_level String Rosé wine subclass-of subclass-of Merlot Chianti 3

4 The Good News Ontologies are the backbone of the Semantic Web More ontologies are available Ontology languages defined as standards: RDF Schema as OWL A huge playing field for ontology research and practice Ontology-development tools lower the barrier for ontology development More people are developing ontologies 4

5 The Ideal World The same language No overlap in coverage No new versions A single extension tree Small reusable modules 5

6 The Bad News: The Real World The same language No overlap in coverage No new versions A single extension tree Small reusable modules 6

7 PROMPT: Dealing with the Messy World Find similarities and differences between ontologies ontology mapping and merging Compare versions of ontologies ontology evolution Extract meaningful portions of ontologies ontology views Integrate in an ontology-editing environment Protégé plugin 7

8 Mapping and Merging Existing ontologies cover overlapping domains use the same terms with different meaning use different terms for the same concept have different definitions for the same concept "Basically, we're all trying to say the same thing." 8

9 iprompt: An Interactive Ontology- Merging and Mapping Tool Declarative mapping iprompt provides Partial automation Algorithm based on concept-representation structure relations between concepts user s actions iprompt does not provide complete automation 9

10 iprompt Algorithm Make initial suggestions Select the next operation Perform automatic updates Find inconsistencies and potential problems Make suggestions 10

11 Example: Merge Classes Activity subclass subclass of of Work Activity subclass subclass of of Meeting Meeting Meeting 11

12 Example: Merge Classes (II) Activity subclass of Meeting attendees attendees present Person Employee 12

13 iprompt: Initial Suggestions 13

14 After a User Performs an Operation For each operation perform the operation consider possible conflicts identify conflicts propose solutions analyze local context create new suggestions reinforce or downgrade existing suggestions 14

15 Analyzing Ontology Structure Structures that Prompt analyzes classes that have the same sets of slots classes that refer to the same set of classes slots that are attached to the same classes Local context incremental analysis consider only the concepts that were affected by the last operation 15

16 AnchorPrompt: Analyzing Graph Structure 16

17 AnchorPrompt: Example Design-a-Trial, S.Modgil, et.al. CMT, I.Sim et.al 17

18 Similarity Score Generate a set of all paths (of length < L) Generate a set of all possible pairs of paths of equal length For each pair of paths and for each pair of nodes in the identical positions in the paths, increment the similarity score Combine the similarity score for all the paths 18

19 AnchorPrompt: Example TRIAL PERSON CROSSOVE Trial Person Crossover PROTOCOL TRIAL-SUBJECT INVESTIGATORS POPULATION PERSON TREATMENT-POPULATION Design Person Person Action_Spec Character Crossover_arm 19

20 AnchorPrompt Discussion Relies on a limited input from the user 3 anchors 2-3 new pairs (above median) 4 anchors 3 new pairs (above median) Has limitations source ontologies with very different structure and level of generality 20

21 Combining Merging and Mapping Declarative mapping 21

22 Prompt Plug-In Architecture Visualization support for comparing concepts Algorithm for initial comparison Presentation of candidate mappings Fine-tuning and saving of mappings Execution of mappings Iterative comparison algorithm 22

23 JambaPrompt: Cognitive Support for Mappings Joint work with Sean Falconer 23

24 The Messy Picture Ontology Change versioning Management 24

25 Comprehensive Ontology Evolution System Editing has subprocess Annotation input produces produces produces Ontology version 1 Changes ChAO refers to Annotation of changes Ontology version 2 input produces input PromptDif algorithm 25

26 Using ChAO ChAO Changes refers to Annotation of changes input input Changes per user/conflicts produces Version comparison input input Accept/reject changes 26

27 ChAO: Change and Annotations Ontology Change applyto author created annotates assoc_annotations Annotation title author created modified Class_Change KB_Change Restriction_Added Class_Created Class_Deleted Superclass_Added 27

28 ChAO Instances in Ontology-Evolution Tasks Examining changes informs the diff display Accepting and rejecting changes stores information on what was accepted Viewing concept history provides access to information for each concept Providing auditing information compiled information on editors and time periods 28

29 Implementation Change-management tab provides access to a list of changes enables annotations Prompt tab generates and presents a diff enables accepting/rejecting changes Core Protégé synchronous editing in a client-server environment undo facilities 29

30 Change-Management Tab 30

31 CMT: Functions Creates ChAO instances in the background provides monitored editing users can examine instances directly in Protégé API access to changes through the Protégé API Provides overview of changes and annotations summary and detailed view of changes annotations for a single change or a group of changes Provides access to concept history access history of changes and annotations from the Classes tab 31

32 PromptDiff Joint work with Michel Klein and Sandhya Kunnatur 32

33 General Problem: Ontology Matching Compare ontologies Find similarities and differences Merging: similarities Mapping: similarities Versioning: differences Ontology Versioning If things look similar, they probably are A large fraction of ontologies remains unchanged from version to version 33

34 The PrompDiff Algorithm Goal: Find a diff automatcally Consists of two parts A set of heuristic matchers A fixed-point algorithm to combine the results of the matchers Can be extended with any number of matchers 34

35 Single Unmatched Siblings Version 1 Version 2 Wine maker Winery color String Wine produced_by Winery White wine Blush wine White wine Rosé wine Red wine Red wine tannin String Merlot Chianti Merlot Chianti 35

36 Siblings with the Same Suffixes or Prefixes Wine maker Winery color String Wine produced_by Winery White Rosé White wine Rosé wine Red Red wine Merlot Chianti Merlot Chianti 36

37 Other Matchers Unmatched superclasses Inverse slots Multiple unmatched siblings Instances of the same class with the same slot values OWL Anonymous classes 37

38 PromptDiff Functions Create a structural diff from instances of ChAO, if present scalability (NCI Thesaurus) automatically, using heuristics, if there are no ChAO instances generate ChAO instances from diff Create user information list of users who edited the ontology for each user number of concepts they edited number of conflicts 38

39 Accept/Reject Changes in Prompt Accept/reject at different levels of granularity an individual change all changes for a class all changes in a subtree all changes from a user or a set of users all non-conflicting changes from a user or a set of users Save and replay accept/reject decisions 39

40 The Messy Picture 40

41 Ontology Views Extract a self-contained subset of an ontology Ensure that all the necessary concepts are defined in the subontology Specify the depth of transitive closure of relations 41

42 Traversal Views Specification of a traversal view A starter concept Relationships to traverse The depth of traversal along each relationship Can find everything related 42

43 Defining a View 43

44 Saving a View Save a view as instances in an ontology Replay the view on a new version Determine if a view is dirty 44

45 Dealing with a Messy World 45

46 Future Directions Mapping and Merging Finding complex mappings Dealing with uncertainty Maintenance during ontology evolution Versioning Integrating with workflow Scalability Views Non-materialized, dynamic views 46

47 "All I'm saying is now is the time to develop the technology to deflect an asteroid" 47