论文信息 - Adding Real Coefficients to Łukasiewicz Logic: An Application to Neural Networks

Adding Real Coefficients to Łukasiewicz Logic: An Application to Neural Networks

In this paper we shall deal with an extension of Łukasiewicz propositional logic obtained by considering scalar multiplication with real numbers, and we focus on the description of its Lindenbaum algebra, i.e., the algebra of truth functions. We show the correspondence between truth tables of such logic and multilayer perceptrons in which the activation function is the truncated identity.

[1] Wan Ahmad Tajuddin Wan Abdullah,et al. The Logic of Neural Networks , 1993 .

[2] Vladik Kreinovich,et al. How to Interpret Neural Networks in Terms of Fuzzy Logic , 2001 .

[3] DANIELE MUNDICI,et al. Averaging the truth-value in Łukasiewicz logic , 1995, Stud Logica.

[4] S. Ovchinnikov. Max-Min Representation of Piecewise Linear Functions , 2000, math/0009026.

[5] Klaus Keimel,et al. Groupes et anneaux réticulés , 1977 .

[6] Antonio di Nola,et al. Łukasiewicz logic and Riesz spaces , 2014, Soft Comput..

[7] A. Zaanen. Riesz Spaces, II , 1983 .

[8] Stanley Burris,et al. A course in universal algebra , 1981, Graduate texts in mathematics.

[9] Antonio di Nola,et al. Riesz MV-algebras and their logic , 2011, EUSFLAT Conf..

[10] Petr Hájek,et al. Handbook of mathematical fuzzy logic , 2011 .

[11] Daniele Mundici,et al. A constructive proof of McNaughton's theorem in infinite-valued logic , 1994, Journal of Symbolic Logic (JSL).

[12] D. Mundici,et al. Algebraic Foundations of Many-Valued Reasoning , 1999 .

[13] Antonio di Nola,et al. Neural Networks And Rational Mcnaughton Functions , 2005, J. Multiple Valued Log. Soft Comput..

[14] Robert McNaughton,et al. A Theorem About Infinite-Valued Sentential Logic , 1951, J. Symb. Log..