| LPMLN is a knowledge representation and reasoning language which conbining Answer Set Programming?ASP?and Markov Logic Network?MLN?.It is capable of nonmonotonic reasoning and uncertainty reasoning.LPMLN is viewed as one of the core reasoning technology in the 863 project,“The key technology and system for open domain knowledge linking,reasoning and retrieving”,due to its ability of handling with inconsistency in logic program.In the project,it is required to explain the LPMLN reasoning consequence.To meet the requirement,this thesis mainly studies an explanation mechanism of LPMLNLN and then designs and implements a LPMLN reasoning explanation system.The main tasks and contributions in this thesis include:?1?Due to LPMLN is developed from ASP,design an explanation model of ASP reasoning as the basis for explanation mechanism of LPMLN reasoning under the circumstances that all of the logic programs in the project is Normal Logic Program?NLP?.Propose the definition of explanation space for NLP and further define ASP reasoning explanation.According to the definitions design the explanation model of ASP reasoning whose core algorithms are explanation space generation algorithm and ASP reasoning explanation generation algorithm.?2?Design an explanation mechanism of LPMLNLN reasoning based on explanation model of ASP reasoning.Propose the definition of LPMLNLN reasoning explanation which is a set of ASP reasoning explanation with probability.The sum of probability in LPMLNLN reasoning explanation is equal to the probability of explanation target.According to the definition design an explanation model of LPMLN reasoning.By grouding and filtrate LPMLN logic program,the model can turn LPMLN reasoning explanation problems into ASP reasoning explanation problems.?3?Design and implement a LPMLN reasoning explanation system based on explanation mechanism of LPMLNLN reasoning.Verify the effectiveness of the mechanism by some application cases. |
PDF Full Text Request