Interactive and Opportunistic Knowledge Acquisition in Case-Based - PowerPoint PPT Presentation

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A, adaptation knowledge formalism source target Sol ( source ) ? Reformulation r Adaptation path Adaptation function A r Adaptation method AM ( source ) Adaptation operator AO r = ( r , A r ) Adaptation error associated to AO r 18

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A, adaptation knowledge formalism r 1 r 2 r 3 source pb 1 pb 2 target Sol ( source ) ? Reformulation r Adaptation path Adaptation function A r Adaptation method AM ( source ) Adaptation operator AO r = ( r , A r ) Adaptation error associated to AO r 18

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A, adaptation knowledge formalism r 1 r 2 r 3 source pb 1 pb 2 target Sol ( source ) ? A r 1 Reformulation r Adaptation path Adaptation function A r Adaptation method AM ( source ) Adaptation operator AO r = ( r , A r ) Adaptation error associated to AO r 18

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A, adaptation knowledge formalism r 1 r 2 r 3 source pb 1 pb 2 target Sol ( source ) Sol ( pb 1 ) ? A r 1 Reformulation r Adaptation path Adaptation function A r Adaptation method AM ( source ) Adaptation operator AO r = ( r , A r ) Adaptation error associated to AO r 18

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A, adaptation knowledge formalism r 1 r 2 r 3 source pb 1 pb 2 target Sol ( source ) Sol ( pb 1 ) Sol ( pb 2 ) ? A r 1 A r 2 Reformulation r Adaptation path Adaptation function A r Adaptation method AM ( source ) Adaptation operator AO r = ( r , A r ) Adaptation error associated to AO r 18

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A, adaptation knowledge formalism r 1 r 2 r 3 source pb 1 pb 2 target AO Sol ( source ) Sol ( pb 1 ) Sol ( pb 2 ) ? A r 1 A r 2 Reformulation r Adaptation path Adaptation function A r Adaptation method AM ( source ) Adaptation operator AO r = ( r , A r ) Adaptation error associated to AO r 18

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A, adaptation knowledge formalism r 1 r 2 r 3 source pb 1 pb 2 target Sol ( source ) Sol ( pb 1 ) Sol ( pb 2 ) ? A r 1 A r 2 A r 3 Reformulation r Adaptation path Adaptation function A r Adaptation method AM ( source ) Adaptation operator AO r = ( r , A r ) Adaptation error associated to AO r 18

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A, adaptation knowledge formalism r 1 r 2 r 3 source pb 1 pb 2 target Sol ( source ) Sol ( pb 1 ) Sol ( pb 2 ) Sol ( target ) A r 1 A r 2 A r 3 Reformulation r Adaptation path Adaptation function A r Adaptation method AM ( source ) Adaptation operator AO r = ( r , A r ) Adaptation error associated to AO r 18

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Knowledge acquisition in I AK A Adaptation process Sol pb 19

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Knowledge acquisition in I AK A Adaptation process Sol Sol ( source ) pb source 19

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Knowledge acquisition in I AK A Adaptation process Sol Sol ( source ) OA ( source ) pb source 19

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Knowledge acquisition in I AK A Adaptation process Sol Sol ( source ) OA ( source ) pb source target 19

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Knowledge acquisition in I AK A Adaptation process Sol g Sol ( target ) OA ( source ) Sol ( source ) OA ( source ) pb source target 19

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Knowledge acquisition in I AK A Interactions and knowledge acquisition process Sol Role of the user g Sol ( target ) OA ( source ) Say if the solution is correct or not Give the correct solution error > ε Give (or verify) the adaptation operator Sol ( source ) OA ( source ) pb source target 19

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Knowledge acquisition in I AK A Interactions and knowledge acquisition process Sol Role of the user g Sol ( target ) OA ( source ) Say if the solution is correct or not Give the correct solution error > ε Give (or verify) the adaptation operator OA ( target ) Sol ( target ) Sol ( source ) OA ( source ) pb source target 19

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Knowledge acquisition in I AK A Change the adaptation method? Sol Role of the user g Sol ( target ) OA ( source ) Say if the solution is correct or not Give the correct solution error > ε Give (or verify) the adaptation operator OA ( target ) Sol ( target ) Sol ( source ) OA ( source ) pb source target 19

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Experiments Parameters studied Size of the case base (up to 20.000 problems) Influence of the tolerance threshold Number of interactions Impact of the decomposition in steps (from 2 to 50 variables) Scope of the operators (adaptation methods) Impact of discontinuities ( C ∞ on R n ) Statistical tests Z-test Wilcoxon 20

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Modelling the domain 2.5 2 1.5 1 2 0.5 0 0 -2 -0.5 z z -1 -4 -1.5 -6 -2 -8 -2.5 -10 10 5 − 10 0 y − 5 0 − 5 x 5 10 − 10 f ht : R 2 → R q x x 2 + y 2 + g ( x , y ) = sin 7 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Introducing a discontinuity 1.4 1.2 1 2 0.8 0 0.6 -2 z z 0.4 -4 0.2 -6 -8 0 -10 10 5 − 10 0 y − 5 0 − 5 x 5 10 − 10 f ht : R 2 → R h ( x , y ) = g ( x , y ) if x 2 + y 2 ≤ 4 q x x 2 + y 2 + g ( x , y ) = sin 7 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Introducing a discontinuity -8 -8.2 -8.4 2 -8.6 0 -8.8 -2 z z -9 -4 -9.2 -6 -8 -9.4 -10 10 5 − 10 0 y − 5 0 − 5 x 5 10 − 10 f ht : R 2 → R f ht ( x , y ) = − 3 − g ( x , y ) if x 2 + y 2 ≤ 4 q x x 2 + y 2 + g ( x , y ) = sin 7 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? A domain with a discontinuity 4 2 0 2 -2 0 -4 -2 z z -6 -4 -8 -6 -10 -8 -10 10 5 − 10 0 y − 5 − 5 0 x 5 10 − 10 f ht : R 2 → R f ht ( x , y ) = − 3 − g ( x , y ) if x 2 + y 2 ≤ 4 q x f ht ( x , y ) = g ( x , y ) if x 2 + y 2 ≤ 4 x 2 + y 2 + g ( x , y ) = sin 7 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? A domain with a discontinuity 4 2 0 2 -2 0 -4 -2 z z -6 -4 -8 -6 -10 -8 -10 10 5 − 10 0 y − 5 − 5 0 x 5 10 − 10 f ht : R 2 → R f ht ( x , y ) = − 3 − g ( x , y ) if x 2 + y 2 ≤ 4 q x f ht ( x , y ) = g ( x , y ) if x 2 + y 2 ≤ 4 x 2 + y 2 + g ( x , y ) = sin 7 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Generating cases 4 2 0 2 -2 0 -4 z z -2 -6 -4 -8 -6 -8 -10 -10 10 5 − 10 0 − 5 y 0 − 5 x 5 10 − 10 f ht : R 2 → R f ht ( x , y ) = − 3 − g ( x , y ) if x 2 + y 2 ≤ 4 q x f ht ( x , y ) = g ( x , y ) if x 2 + y 2 ≤ 4 x 2 + y 2 + g ( x , y ) = sin 7 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? The system knowledge 2 0 z z -2 -4 -6 -8 -10 10 5 − 10 0 − 5 y 0 − 5 x 5 10 − 10 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Random generation of problems, solved by the system 2 0 z z -2 -4 -6 -8 -10 10 5 − 10 0 − 5 y − 5 0 x 5 10 − 10 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? Problems triggering a knowledge acquisition process... 2 0 z z -2 -4 -6 -8 -10 10 5 − 10 0 − 5 y 0 − 5 x 5 10 − 10 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? ...are located around the discontinuity zone... 4 2 0 2 -2 0 -4 -2 z z -6 -4 -8 -6 -8 -10 -10 10 5 − 10 0 y − 5 0 − 5 x 5 10 − 10 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? ...allowing the identification of a critical part of the domain 4 2 0 2 -2 0 -4 z z -2 -6 -4 -8 -6 -8 -10 -10 10 5 − 10 0 − 5 y − 5 0 x 5 10 − 10 21

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? I AK A contributions I AK A: Main contributions Formalisation of the adaptation knowledge Adaptation operators ( AO r = ( r , A r ) ) Adaptation methods ( AM ( source ) ) Adaptation path, adaptation step Failure-driven knowledge acquisition strategy Focus on adaptation knowledge acquisition Decomposition into steps to easily identify faulty knowledge Prototype implementing the differential adaptation strategy Numerical functions formalism Differential operators Experiments Test of the method with several parameters: expert tolerance threshold, number of cases, number of interactions... Experiments with CBR hypotheses, “Similar problems have similar solutions” 22

Intro FIKA I AK A F RAKA S T AAABLE What’s next? F RAKA S FailuRe-driven interactive Adaptation Knowledge AcquiSition FIKA I AK A F RAKA S T AAABLE 23

Intro FIKA I AK A F RAKA S T AAABLE What’s next? F RAKA S, an interactive domain knowledge acquisition approach F RAKA S properties FIKA Knowledge-intensive Interactive Opportunistic On-line User-centric 24

Intro FIKA I AK A F RAKA S T AAABLE What’s next? F RAKA S, an interactive domain knowledge acquisition approach F RAKA S properties FIKA Knowledge-intensive I AK A F RAKA S Interactive Opportunistic On-line User-centric 24

Intro FIKA I AK A F RAKA S T AAABLE What’s next? F RAKA S, an interactive domain knowledge acquisition approach F RAKA S properties FIKA Knowledge-intensive I AK A F RAKA S Interactive Opportunistic F RAKA S-PL On-line User-centric 24

Intro FIKA I AK A F RAKA S T AAABLE What’s next? F RAKA S, an interactive domain knowledge acquisition approach F RAKA S properties FIKA Knowledge-intensive I AK A F RAKA S Interactive Opportunistic F RAKA S-PL On-line F RAKA S-PL(O NCO ) User-centric 24

Intro FIKA I AK A F RAKA S T AAABLE What’s next? From belief revision theory to conservative adaptation Belief revision: updating a knowledge base while maintaining consistency Minimal change: revision operator makes a minimal change on the initial base µ ψ ψ ◦ µ Conservative adaptation CA ◦ ( SDK , source ∧ Sol ( source ) , target ) = ( SDK ∧ source ∧ Sol ( source )) ◦ ( SDK ∧ target ) 25

Intro FIKA I AK A F RAKA S T AAABLE What’s next? From belief revision theory to conservative adaptation Belief revision: updating a knowledge base while maintaining consistency Minimal change: revision operator makes a minimal change on the initial base ◦ µ ψ ψ ◦ µ Conservative adaptation CA ◦ ( SDK , source ∧ Sol ( source ) , target ) = ( SDK ∧ source ∧ Sol ( source )) ◦ ( SDK ∧ target ) 25

Intro FIKA I AK A F RAKA S T AAABLE What’s next? From belief revision theory to conservative adaptation Belief revision: updating a knowledge base while maintaining consistency Minimal change: revision operator makes a minimal change on the initial base ◦ Conservative adaptation CA ◦ ( SDK , source ∧ Sol ( source ) , target ) = ( SDK ∧ source ∧ Sol ( source )) ◦ ( SDK ∧ target ) 25

Intro FIKA I AK A F RAKA S T AAABLE What’s next? From belief revision theory to conservative adaptation Belief revision: updating a knowledge base while maintaining consistency Minimal change: revision operator makes a minimal change on the initial base target target source Sol ( source ) Sol ( target ) ◦ SDK SDK SDK Conservative adaptation CA ◦ ( SDK , source ∧ Sol ( source ) , target ) = ( SDK ∧ source ∧ Sol ( source )) ◦ ( SDK ∧ target ) 25