1 / 14

WP03 AI tools: Case Based Reasoning

WP03 AI tools: Case Based Reasoning. Miquel Sànchez-Marrè Knowledge Engineering & Machine Learning Group http://www.lsi.upc.es/ webia/keml Dept. Of Computer Software Technical University of Catalonia (UPC). WP03 objectives. Creation of a web-accesible Case-Based Reasoning tool.

ewan
Télécharger la présentation

WP03 AI tools: Case Based Reasoning

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. WP03AI tools: Case Based Reasoning Miquel Sànchez-Marrè Knowledge Engineering & Machine Learning Group http://www.lsi.upc.es/webia/keml Dept. Of Computer Software Technical University of Catalonia (UPC)

  2. WP03 objectives • Creation of a web-accesible Case-Based Reasoning tool. • Integration of Case-Based Reasoning into the Emergency Management Trainer: • Integration of CBR tools with the primary emergency simulation, that augments the simulation tools with an empirical approach based on similarity and analogy; • Integration of CBR concepts and tools in the didactic components of the system for the analysis of student performance and the selection of illustrative examples for in-depth explanation.

  3. DONE In progress WP03 work plan • Creation of the CBR server • D03.1 AI tools: CBR (Implementation report) • D03.2 AI tools: CBR (Prototype implementation) • Connection with the RTXPS server • Introduction of cases in the Case Library

  4. CBR: A methodology of solving new problems by adapting the solutions of previous similar problems It uses cases as an episodic memory (Case Library). Case-Based Reasoning (I) Solution New Solution Problem New Problem

  5. Case-Based Reasoning (II) • A case is a set of features • case identifier • derivation of the case • description of the problem • diagnostic of the problem • solution to the problem • evaluation of the solution (success/failure) • utility measure • other relevant information

  6. CASE LIBRARY Case-Based Reasoning (III) • The CBR cycle: new case Retrieve Learn retrieved cases case to store best case evaluated solution (fail/ success) DOMAIN KNOWLEDGE Adapt adapted solution Eval

  7. CASE LIBRARY How is the A-TEAM CBR? (I) CBR module new case emergency Retrieve retrieved cases best case DOMAIN KNOWLEDGE adapted solution Adapt simulation

  8. How is the A-TEAM CBR? (II) • Retrieval: • Retrieving cases is more difficult than retrieval in DB. • Retrieval in DB = exact matching • Retrieval in CBR = partial matching (similarity) • Similarity: • is computed among case descriptions, • is usually an evaluation heuristic function or distance, • can be domain-dependent. • A-TEAM approach: • case description = set of attributes • similarity measure = • Retrieval maximises the similarity between the actual case and the retrieved one(s).

  9. How is the A-TEAM CBR? (III) • Adaptation: • When the case selected from the Case Lib. does not match perfectly with the new case, the old solution needs to be adapted to the new case solution. • Strategies: • null adaptation • structural adaptation • substitution methods • transformation methods • special-purpose adaptation • derivational adaptation • Adaptation is high case-study dependent.

  10. Meta-data case structure in A-TEAM

  11. The CBR server: Interactions response (simulation) • Comunication among RTXPS and CBR through HTTP POST Method CBR Server RTXPS Server Simulation Module simulation data request (emergency) off-line on-line

  12. CASE LIBRARY The CBR server: parts CBR Server query HTTP request CBR core HTTP server HTTP response solution Http connections Socket connections

  13. The CBR server: serving a request (I) HTTP Server HTTP request RTXPS server Web Server HTTP response CBR core POST parameters Results (text) query Java Servlet solution

  14. The CBR server: serving a request (II) RTXPS Parameter Descriptors CBR core Case Library Builder Simulated Emergencies Data CASE LIBRARY HTTP server query Case Library Manager solution

More Related