Adaptive Reasoning for Cooperative Agents
Using explicit affirmation and explicit negation, whilst allowing for a third logic value of undefinedness, can be useful in situations where decisions have to be taken on the basis of scarce, ambiguous, or downright contradictory information. In a three-valued setting, we consider an agent that learns a definition for both the target concept and its opposite, considering positive and negative examples as instances of two disjoint classes. Explicit negation is used to represent the opposite concept, while default negation is used to ensure consistency and to handle exceptions to general rules. Exceptions are represented by examples covered by the definition for a concept that belong to the training set for the opposite concept.
One single agent exploring an environment may gather only so much information about it and that may not suffice to find the right explanations. In such case, a cooperative multi-agent strategy, where each agent explores a part of the environment and shares with the others its findings, might provide better results. We describe one such framework based on a distributed genetic algorithm enhanced by a Lamarckian operator for belief revision. The agents communicate their candidate explanations — coded as chromosomes of beliefs — by sharing them in a common pool. Another way of interpreting this communication is in the context of argumentation. In the process of taking all the arguments and trying to find a common ground or consensus we might have to change, or review, some of assumptions of each argument.
The resulting framework we present is a collaborative perspective of argumentation where arguments are put together at work in order to find the possible 2-valued consensus of opposing positions of learnt concepts in an evolutionary pool in order to find the “best” explanation to the observations.
Unable to display preview. Download preview PDF.
- 2.Alferes, J.J., Damásio, C.V., Pereira, L.M.: SLX - A top-down derivation procedure for programs with explicit negation. In: Bruynooghe, M. (ed.) Proc. Int. Symp. on Logic Programming. The MIT Press, Cambridge (1994)Google Scholar
- 9.De Raedt, L., Bruynooghe, M.: Towards friendly concept-learners. In: Procs. of the 11th Intl. Joint Conf. on Artificial Intelligence, pp. 849–856. Morgan Kaufmann, San Francisco (1989)Google Scholar
- 10.De Raedt, L., Bruynooghe, M.: On negation and three-valued logic in interactive concept learning. In: Procs. of the 9th European Conf. on Artificial Intelligence (1990)Google Scholar
- 13.Esposito, F., Ferilli, S., Lamma, E., Mello, P., Milano, M., Riguzzi, F., Semeraro, G.: Cooperation of abduction and induction in logic programming. In: Flach, P.A., Kakas, A.C. (eds.) Abductive and Inductive Reasoning, Pure and Applied Logic. Kluwer, Dordrecht (1998)Google Scholar
- 14.Gelfond, M., Lifschitz, V.: The stable model semantics for logic programming. In: Kowalski, R., Bowen, K.A. (eds.) Procs. of the 5th Int. Conf. on Logic Programming, pp. 1070–1080. MIT Press, Cambridge (1988)Google Scholar
- 16.Green, D.M., Swets, J.M.: Signal detection theory and psychophysics. John Wiley and Sons Inc., New York (1966) ISBN 0-471-32420-5Google Scholar
- 17.Greiner, R., Grove, A.J., Roth, D.: Learning active classifiers. In: Procs. of the Thirteenth Intl. Conf. on Machine Learning, ICML 1996 (1996)Google Scholar
- 18.Inoue, K.: Learning abductive and nonmonotonic logic programs. In: Flach, P.A., Kakas, A.C. (eds.) Abductive and Inductive Reasoning, Pure and Applied Logic. Kluwer, Dordrecht (1998)Google Scholar
- 19.Inoue, K., Kudoh, Y.: Learning extended logic programs. In: Procs. of the 15th Intl. Joint Conf. on Artificial Intelligence, pp. 176–181. Morgan Kaufmann, San Francisco (1997)Google Scholar
- 21.Lamarck, J.B.: http://www.ucmp.berkeley.edu/history/lamarck.html
- 22.Grefenstette, J.J.: Lamarckian learning in multi-agent environments (1991)Google Scholar
- 23.Lamma, E., Riguzzi, F., Pereira, L.M.: Learning in a three-valued setting. In: Procs. of the Fourth Intl. Workshop on Multistrategy Learning (1988)Google Scholar
- 26.Lapointe, S., Matwin, S.: Sub-unification: A tool for efficient induction of recursive programs. In: Sleeman, D., Edwards, P. (eds.) Procs. of the 9th Intl. Workshop on Machine Learning, pp. 273–281. Morgan Kaufmann, San Francisco (1992)Google Scholar
- 27.Lavrač, N., Džeroski, S.: Inductive Logic Programming: Techniques and Applications. Ellis Horwood (1994)Google Scholar
- 28.Pazzani, M.J., Merz, C., Murphy, P., Ali, K., Hume, T., Brunk, C.: Reducing misclassification costs. In: Procs. of the Eleventh Intl. Conf. on Machine Learning (ML 1994), pp. 217–225 (1994)Google Scholar
- 30.Pereira, L.M., Pinto, A.M.: Reductio ad absurdum argumentation in normal logic programs. In: Argumentation and Non-monotonic Reasoning (ArgNMR 2007) Workshop at LPNMR 2007, pp. 96–113 (2007)Google Scholar
- 32.Pinto, A.M.: Explorations in revised stable models — a new semantics for logic programs. Master’s thesis, Universidade Nova de Lisboa (February 2005)Google Scholar
- 33.Provost, F.J., Fawcett, T.: Analysis and visualization of classifier performance: Comparison under imprecise class and cost distribution. In: Procs. of the Third Intl. Conf. on Knowledge Discovery and Data Mining (KDD 1997). AAAI Press, Menlo Park (1997)Google Scholar
- 34.Quinlan, J.: Learning logical definitions from relations. Machine Learning 5, 239–266 (1990)Google Scholar
- 35.Vere, S.A.: Induction of concepts in the predicate calculus. In: Procs. of the Fourth Intl. Joint Conf. on Artificial Intelligence (IJCAI 1975), pp. 281–287 (1975)Google Scholar
- 36.Whitley, D., Rana, S., Heckendorn, R.B.: The Island Model Genetic Algorithm: On Separability, Population Size and Convergence (1998)Google Scholar