Abstract
This paper is a plea for developing possibilistic learning methods that would be consistent with if-then rule-based reasoning. The paper first recall the possibility theory-based handling of cascading sets of parallel if-then rules. This is illustrated by an example describing a classification problem. It is shown that the approach is both close to a possibilistic logic handling of the problem and can also be put under the form of a max-min-based matrix calculus describing a function underlying a structure somewhat similar to a max-min neural network. The second part of the paper discusses how possibility distributions can be obtained from precise or imprecise statistical data, and then surveys the few existing works on learning in a possibilistic setting. A final discussion emphasizes the interest of handling learning and reasoning in a consistent way.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
For the \(\max \)-\(\min \) composition, \(\begin{bmatrix} \pi (q) \\ \pi (\lnot q) \end{bmatrix} = \begin{bmatrix} \varPi (p \wedge q) &{} \ \varPi (\lnot p \wedge q) \\ \varPi (p \wedge \lnot q) &{} \ \varPi (\lnot p\wedge \lnot q) \end{bmatrix} \otimes \begin{bmatrix} \pi (p) \\ \pi (\lnot p)\end{bmatrix} \ \ \ (1')\), taking advantage of the monotonicity and of the max-decomposability of possibility measures. In contrast with (1), there are inequality constraints between matrix terms and \((\pi (p), \pi (\lnot p))\).
- 2.
By duality the inequality (2) writes \(\begin{bmatrix} \pi (q) \\ \pi (\lnot q) \end{bmatrix} \le \begin{bmatrix} \varPi (p \wedge q) &{} \ \varPi (\lnot p \wedge q) \\ \varPi (p \wedge \lnot q) &{} \ \varPi (\lnot p \wedge \lnot q) \end{bmatrix} \boxdot \begin{bmatrix} \pi (\lnot p) \\ \pi (p)\end{bmatrix}\ \ \ (2')\) where \(\boxdot \) is the \(\min -\max \) product. Note that (2’) provides an upper bound for \(\begin{bmatrix} \pi (q) \\ \pi (\lnot q) \end{bmatrix}\) obtained by (1’). It can be checked that this upper bound coincides with this latter vector if \(\varPi (p \wedge q) = \varPi (\lnot p \wedge \lnot q)=1\) or if \(\varPi (\lnot p \wedge q)=\varPi (p \wedge \lnot q)=1\).
References
Benferhat, S., Dubois, D., Garcia, L., Prade, H.: On the transformation between possibilistic logic bases and possibilistic causal networks. Int. J. Approx. Reas. 29(2), 135–173 (2002)
Benferhat, S., Dubois, D., Lagrue, S., Prade, H.: A big-stepped probability approach for discovering default rules. Int. J. Uncert. Fuzz. Knowl.-Based Syst. 11(Suppl. 1), 1–14 (2003)
Benferhat, S., Dubois, D., Prade, H.: Possibilistic and standard probabilistic semantics of conditional knowledge bases. J. Log. Comput. 9(6), 873–895 (1999)
Benferhat, S., Tabia, K.: Binary naive possibilistic classifiers: handling uncertain inputs. Int. J. Intell. Syst. 24(12), 1203–1229 (2009)
Borgelt, C., Kruse, R.: Learning from imprecise data: possibilistic graphical models. Comput. Stat. Data Anal. 38(4), 449–463 (2002)
Borgelt, C., Steinbrecher, M., Kruse, R.: Graphical Models: Representations for Learning, Reasoning and Data Mining, 2nd edn. Wiley (2009)
Bounhas, M., Ghasemi Hamed, M., Prade, H., Serrurier, M., Mellouli, K.: Naive possibilistic classifiers for imprecise or uncertain numerical data. Fuzzy Sets Syst. 239, 137–156 (2014)
Bouraoui, Z., et al.: From shallow to deep interactions between knowledge representation, reasoning and machine learning (Kay R. Amel group). CoRR abs/1912.06612 (2019)
Buckley, J.J., Hayashi, Y.: Fuzzy neural networks: a survey. Fuzzy Sets Syst. 66, 1–13 (1994)
Delgado, M., Moral, S.: On the concept of possibility-probability consistency. Fuzzy Sets Syst. 21, 311–318 (1987)
Denœux, T.: Logistic regression, neural networks and Dempster-Shafer theory: a new perspective. Knowl.-Based Syst. 176, 54–67 (2019)
Dubois, D., Foulloy, L., Mauris, G., Prade, H.: Probability-possibility transformations, triangular fuzzy sets, and probabilistic inequalities. Reliable Comput. 10, 273–297 (2004)
Dubois, D., Prade, H.: On several representations of an uncertain body of evidence. In: Gupta, M., Sanchez, E. (eds.) Fuzzy Information and Decision Processes, pp. 167–181. North-Holland, Amsterdam (1982)
Dubois, D., Prade, H.: Unfair coins and necessity measures: a possibilistic interpretation of histograms. Fuzzy Sets Syst. 10(1), 15–20 (1983)
Dubois, D., Prade, H.: Fuzzy sets and statistical data. Europ. J. Oper. Res. 25, 345–356 (1986)
Dubois, D., Prade, H.: Possibility Theory - An Approach to Computerized Processing of Uncertainty. Plenum Press (1988)
Dubois, D., Prade, H.: Consonant approximations of belief functions. Int. J. Approximate Reasoning 4, 419–449 (1990)
Dubois, D., Prade, H.: On the validation of fuzzy knowledge bases. In: Tzafestas, S.G., Venetsanopoulos, A.N. (eds.) Fuzzy Reasoning in Information, Decision and Control Systems, pp. 31–49. Kluwer Acad. Publ. (1994)
Dubois, D., Prade, H.: Fuzzy criteria and fuzzy rules in subjective evaluation - a general discussion. In: Proceedings of the 5th European Congress on Intelligent Techniques and Soft Computing (EUFIT 1997), Aachen, vol. 1, pp. 975–979 (1997)
Dubois, D., Prade, H., Yager, R.R.: Merging fuzzy information. In: Bezdek, J., Dubois, D., Prade, H. (eds.) Fuzzy Sets in Approximate Reasoning and Information Systems. The Handbooks of Fuzzy Sets Series, pp. 335–401. Kluwer, Boston (1999)
Dubois, D., Prade, H.: An overview of ordinal and numerical approaches to causal diagnostic problem solving. In: Gabbay, D.M., Kruse, R. (eds.) Abductive Reasoning and Uncertainty Management Systems. Handbook of Defeasible Reasoning and Uncertainty Management Systems, vol. 4, pp. 231–280. Kluwer Acad. Publ. (2000). Gabbay, D.M., Smets, Ph. (series eds.)
Dubois, D., Prade, H.: Possibility theory and its applications: where do we stand? In: Kacprzyk, J., Pedrycz, W. (eds.) Springer Handbook of Computational Intelligence, pp. 31–60. Springer, Heidelberg (2015). https://doi.org/10.1007/978-3-662-43505-2_3
Dubois, D., Prade, H.: Practical methods for constructing possibility distributions. Int. J. Intell. Syst. 31(3), 215–239 (2016)
Dubois, D., Prade, H.: Towards a logic-based view of some approaches to classification tasks. In: Lesot, M.-J., et al. (eds.) IPMU 2020. CCIS, vol. 1239, pp. 697–711. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-50153-2_51
Dubois, D., Prade, H., Rico, A.: The logical encoding of Sugeno integrals. Fuzzy Sets Syst. 241, 61–75 (2014)
Dubois, D., Prade, H., Schockaert, S.: Generalized possibilistic logic: foundations and applications to qualitative reasoning about uncertainty. Artif. Intell. 252, 139–174 (2017)
Dubois, D., Prade, H., Smets, P.: A definition of subjective possibility. Int. J. Approx. Reasoning 48(2), 352–364 (2008)
Farreny, H., Prade, H.: Default and inexact reasoning with possibility degrees. IEEE Trans. Syst. Man Cybern. 16(2), 270–276 (1986)
Farreny, H., Prade, H.: Positive and negative explanations of uncertain reasoning in the framework of possibility theory. In: Proceedings of the 5th Conference on Uncertainty in Artificial Intelligence (UAI 1989), Windsor, ON, August 18–20, pp. 95–101 (1989). Expanded version: explications de raisonnements dans l’incertain. Rev. d’Intell. Artif. 4(2), 43–75 (1990)
Farreny, H., Prade, H., Wyss, E.: Approximate reasoning in a rule-based expert system using possibility theory: a case study. In: Kugler, H.J. (ed.) Proceedings of the 10th World IFIP Congress on Information Processing 1986, Dublin, September 1–5, pp. 407–413. North-Holland (1986)
Gebhardt, J., Kruse, R.: Learning possibilistic networks from data. In: Fisher, D., Lenz, H.-J. (eds.) Learning from Data - Proceedings of the 5th International Workshop on Artificial Intelligence and Statistics, Key West, January 1995, pp. 143–153. Springer, New York (1996)
Green, T.J., Karvounarakis, G., Tannen, V.: Provenance semirings. In: Libkin, L. (ed.) Proceedings of the 26th ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems, Beijing, June 11–13, pp. 31–40 (2007)
Haddad, M., Leray, P., Ben Amor, N.: Apprentissage des réseaux possibilistes à partir de données: un survol. Rev. d’Intell. Artif. 29(2), 229–252 (2015)
Haddad, M., Leray, Ph., Ben Amor, N.: Possibilistic networks: parameters learning from imprecise data and evaluation strategy. arXiv:1607.03705 [cs.AI] (2016)
Haddad, M., Leray, Ph., Ben Amor, N.: Possibilistic MDL: a new possibilistic likelihood based score function for imprecise data. In: ECSQARU 2017, pp. 435–445 (2017)
Haddad, M., Leray, Ph., Levray, A., Tabia, K.: Learning the parameters of possibilistic networks from data: empirical comparison. In: Proceedings of the FLAIRS Conference, pp. 736–741 (2017)
Jenhani, I., Ben Amor, N., Benferhat, S., Elouedi, Z.: SIM-PDT: a similarity based possibilistic decision tree approach. In: Hartmann, S., Kern-Isberner, G. (eds.) FoIKS 2008. LNCS, vol. 4932, pp. 348–364. Springer, Heidelberg (2008). https://doi.org/10.1007/978-3-540-77684-0_23
Kampé de Fériet, J.: Interpretation of membership functions of fuzzy sets in terms of plausibility and belief. In: Gupta, M., Sanchez, E. (eds.) Fuzzy Information and Decision Processes, pp. 93–98. North-Holland, Amsterdam (1982)
Kraus, S., Lehmann, D., Magidor, M.: Nonmonotonic reasoning, preferential models and cumulative logics. Artif. Intell. 44, 167–207 (1990)
Kuzelka, O., Davis, J., Schockaert, S.: Encoding Markov logic networks in possibilistic logic. In: Meila, M., Heskes, T. (eds.) Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence (UAI 2015), Amsterdam, July 12–16, pp. 454–463. AUAI Press (2015)
Kuzelka, O., Davis, J., Schockaert, S.: Induction of interpretable possibilistic logic theories from relational data. In: Proceedings of the IJCAI 2017, Melbourne, pp. 1153–1159 (2017)
Labreuche, C.: A general framework for explaining the results of a multi-attribute preference model. Artif. Intell. 175(7–8), 1410–1448 (2011)
Mamdani, E.H., Assilian, S.: An experiment in linguistic synthesis with a fuzzy logic controller. Int. J. Man-Mach. Stud. 7(1), 1–13 (1975)
Prade, H., Rico, A., Serrurier, M.: Elicitation of sugeno integrals: a version space learning perspective. In: Rauch, J., Raś, Z.W., Berka, P., Elomaa, T. (eds.) ISMIS 2009. LNCS (LNAI), vol. 5722, pp. 392–401. Springer, Heidelberg (2009). https://doi.org/10.1007/978-3-642-04125-9_42
Richardson, M., Domingos, P.M.: Markov logic networks. Mach. Learn. 62(1–2), 107–136 (2006)
Serrurier, M., Prade, H.: Introducing possibilistic logic in ILP for dealing with exceptions. Artif. Intell. 171(16–17), 939–950 (2007)
Serrurier, M., Prade, H.: An informational distance for estimating the faithfulness of a possibility distribution, viewed as a family of probability distributions, with respect to data. Int. J. Approx. Reason. 54(7), 919–933 (2013)
Serrurier, M., Prade, H.: Entropy evaluation based on confidence intervals of frequency estimates: application to the learning of decision trees. In: Proceedings of the 32nd International Conference on Machine Learning, Lille, July 6–11, pp. 1576–1584 (2015)
Shafer, G.: A Mathematical Theory of Evidence. Princeton University Press (1976)
Sudkamp, T.: On probability-possibility transformations. Fuzzy Sets Syst. 51, 73–81 (1992)
Teow, L.-N., Loe, K.-F.: An effective learning method for max-min neural networks. In: Proceedings of the IJCAI 1997, Nagoya, pp. 1134–1139 (1997)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Dubois, D., Prade, H. (2020). From Possibilistic Rule-Based Systems to Machine Learning - A Discussion Paper. In: Davis, J., Tabia, K. (eds) Scalable Uncertainty Management. SUM 2020. Lecture Notes in Computer Science(), vol 12322. Springer, Cham. https://doi.org/10.1007/978-3-030-58449-8_3
Download citation
DOI: https://doi.org/10.1007/978-3-030-58449-8_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-58448-1
Online ISBN: 978-3-030-58449-8
eBook Packages: Computer ScienceComputer Science (R0)