1 / 41

Strategies for Distributed CBR

Strategies for Distributed CBR. Santi Ontañón IIIA-CSIC. OUTLINE. CBR Introduction Applications Distributed CBR Autonomous CBR agents Collaboration policies Learning to collaborate Case Bartering Conclusions. What is CBR?.

tad-simpson
Download Presentation

Strategies for Distributed CBR

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. Strategies for Distributed CBR Santi Ontañón IIIA-CSIC

  2. OUTLINE • CBR • Introduction • Applications • Distributed CBR • Autonomous CBR agents • Collaboration policies • Learning to collaborate • Case Bartering • Conclusions

  3. What is CBR? • Case Based Reasoning: To solve a new problem by noticing its similarity to one or several previously solved problems and by adapting their solutions instead of working out a solution from scratch.

  4. CBR: Cycle New case PROBLEM RETRIEVE General Knowledge CASE BASE RETAIN Tested Case Retrieved cases REUSE REVISE Solved case

  5. CBR: Retrieve • Is the key problem in CBR • Indexing: • Relevant features: predictive, discriminatory and explanatory. • Good case representation • Similarity metrics

  6. CBR: Reuse • Copy solution (for classification) • Modify old solution: • Reinstantiate (variables) • Local search • Transform • Domain knowledge (causal model)

  7. CBR: Revise • Apply solution to problem: • In real world • In simulated world • Evaluate • Repair • User guided • Internal (Domain knowledge)

  8. CBR: Retain • What to retain? • Relevant problems • Solution (failed/successful) • Solution method • Store case • Update/indentify indexes

  9. CBR systems: • CHEF, cooking planing (Hammon 86) • HYPO, legal reasoning (Ashley/Rissland 87) • PROTOS, medical diagnosis (Bereiss/Porter 88) • CASEY, medial diagnosis (Koton 89) • SAXEX, expresive music (Arcos 96)

  10. Distributed CBR • Autonomous CBR agents • Collaboration policies • Learning to collaborate • Case Bartering

  11. Autonomous CBR agents • Multiagent CBR system • Each agent has an individual Case-Base • Agent has autonomy in • Problem acquisition • Problem solving • Learning • Collaboration policies • Explicit strategies to improve individual performance

  12. Multiagent CBR system • MAC (Multiagent CBR) System • An agent is a pair (Ai,Ci) • A Case base • Classification task in

  13. Collaboration policies • Committee • Peer Counsel • Bounded Counsel

  14. Collaboration • An agent A1 with problem P • Sends P to some agents A={ A2,…,An } • Each Aj answers with a Solution Endorsing Record (SER) • < { (S1 Ej1)…(Sm Ejm) } , P , Aj > • Endorsing pairs: ( Sk,Ejk )

  15. Solution EndorsingRecord (SER) P P1 P2 P3 P4 S1 S1 S2 S3 { (S2,1), (S3,2) } P5 P6 S2 P7 S2 P8 S3 S1 P9 S4 Relevant cases

  16. Voting(1) • At is the set of agents submitting SER at time t (including the initiating agent). • Each agent has 1 vote • The vote can be fractionally assigned to various solution classes:

  17. Voting(2) • The agent that receives de SERs, aggregates the votes: • The proposed solution is that with maximum Ballot:

  18. Committee policy (1) • An agent A1 with problem P • Sends P to all the other agents A={ A2,…,An } • Each Aj answers with a Solution Endorsing Record (SER) • < { (S1 Ej1)…(Sm Ejm) } , P , Aj > • The majority vote selects the solution class

  19. Agent interaction: Committee A1 A2 C1 C2 P A3 C3

  20. Agent interaction: Committee A1 A2 C1 C2 P P P A3 C3

  21. Agent interaction: Committee A1 A2 C1 C2 SER of A2 P SER of A1 A3 C3 SER of A3

  22. Agent interaction: Committee A1 A2 C1 C2 P SER of A1 SER of A2 SER of A3 A3 C3 Voting SOLUTION

  23. Committee policy (2) • Robust • Good accuracy even with big number of agents with small case-bases • Cost • Linear with number of agents • Always asks all agents

  24. Peer Counsel • The agent try to solve the problem by itself • Assess competence of current solution: • Votes for the max class >> rest of votes: • If not competent, ask all other agents • Like committee

  25. Agent interaction:Peer Counsel A1 A2 C1 C2 P A3 C3

  26. Agent interaction:Peer Counsel A1 A2 C1 C2 P SER of A1 Competent! A3 C3

  27. Agent interaction:Peer Counsel A1 A2 C1 C2 P SER of A1 A3 Voting C3 SOLUTION

  28. Bounded Counsel • Use self-competence model • If not competent, ask one agent • Agent returns a SER • Assess competence of current solution (Termination Check): • Votes for the max class >> rest of votes: • If not competent, ask one agent more

  29. Agent interaction:Bounded Counsel A1 A2 C1 C2 P A3 C3

  30. Agent interaction:Bounded Counsel A1 A2 C1 C2 P SER of A1 Not competent! A3 C3

  31. Agent interaction:Bounded Counsel A1 A2 C1 C2 P P SER of A1 A3 C3

  32. Agent interaction:Bounded Counsel A1 A2 C1 C2 SER of A2 P SER of A1 A3 C3

  33. Agent interaction:Bounded Counsel A1 A2 C1 C2 P SER of A1 SER of A2 Competent! A3 C3

  34. Agent interaction:Bounded Counsel A1 A2 C1 C2 P SER of A1 SER of A2 A3 Voting C3 SOLUTION

  35. Learning to collaborate • Bounded Counsel • fixed predicate to assess competence • Proactive Learning • Improve collaboration • Actions performed in order to learn • learning when to ask counsel

  36. Learning to collaborate(2) Do I need the counsel of another agent? A1 C1 P SER of A1 Voting SER of A2 CANDIDATE SOLUTION SER of A3

  37. Case Bartering • Why? • Biased individual case-bases • How? • Estimating underlying distribution • Interchanging cases between agents

  38. Case Bartering(2) • Estimation of underlying distribution: • Aggregating statistics of all the individual case-bases.

  39. Case Bartering(3) • Case-Base bias: • The agents should minimize its bias

  40. Case Bartering(4) • Bartering Protocol: • Estimate underlying distribution • Send offers • Confirm offers • Interchange cases • If changes go to 2, else END.

  41. Conclusions • Multi-agent systems offer new paradigms to organize AI. • Agent interaction is still a stimulating challenge. • CBR fits perfectly with these kind of applications.

More Related