Abstract
In this lecture we present a short introduction to the theory of neural networks, in which the mathematical aspects are given special consideration. Static and dynamic aspects of models for pattern retrieval are discussed as well as a statistical mechanical treatment of the Hopfield model in the case where the number of patterns and neurons is large. As a motivation for the study of neural networks we mention the contrast between the fact that von Neumann’s type of sequential or parallel computers are quick in algorithmic computations but slow in performing recognition of patterns, whereas brains (not necessarily human brains) are comparatively slow in algorithmic computations but much quicker in performing pattern recognition. Neural networks (as “caricatures” of biological neuronal networks) are attempts to understand operationally such features of quick recognition of patterns (or signals). In fact they show some efficiency in performing “associative, adaptive memory” (as manifested in retrieval of memories and recognition of simple patterns). The methods used for their study stem from statistical mechanics and the theory of stochastic processes. This is particularly evident in the study of the Hopfield model (Hopfield 1982) to which we shall dedicate most attention in these lectures.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Amit D. (1989): Modeling brain function, Cambridge Univ. Press.
Albeverio S., Feng J. and Tirozzi B. (1997): Rigorous results in neural networks, book in preperation.
Albeverio S., Kondratiev Yu. and Röckner M. (1997): J. Funct. Anal. 144, 394–423.
Arnold V.I. and Avez A. (1968): Ergodic Problems of Classical Mechanics, Benjamin, New York.
Albeverio S., Tirozzi B. and Zegarlinski B. (1992): Comm. Math. Phys. 150, 337–373.
Ay N. (1996): Ein kombinatorisch geometrischer Zugang zum Hopfield-Modell, Diplomarbeit, Bochum.
Baxter R.J. (1982): Exactly solvable models in statistical mechanics, Acadamic Press, London.
Bovier A. and Gayrard V. (1992): J. Stat. Phys. 69, 597–627.
Bovier A. and Gayrard V. (1997): On the retrieval states of the Hopfield Model, Prob. Theory. relat. Fields to appear.
Bovier A. and Gayrard V. (1995): An Almost Sure Large Deviation Principle for the Hopfield Model, Preprint 146, Weierstrasse Institute, Berlin.
Bovier A., Gayrard V. and Picco P. (1994): Probab. Theory Related Fields 100, 329–363.
Feng J. and Tirozzi B. (1997): Capacity of the Hopfield Model J. of Physics A in press.
Guerra F. (1995): About the overlap distribution in mean field spin glass models preprint.
Hopfield J.J. (1982): Proc. Natl. Acad. Sci. USA 79, 1554–2558.
Hecht-Nielsen R. (1990): Neurocomputing, Addison-Wesley, Reading.
losifescu M. (1980): Finite Markov Processes and their application, Wily, Chiherter.
Kinnebrock W. (1992): Neuronale Netze, Oldenbourg, München.
Loukianova D. (1994): Two rigorous bounds in the Hopfield model of associative memory, Manuscript.
Loukianova D. (1994): Capacité de mémoire dlans le modèle de Hopfield, C. R. Acad. Sci. Paris, 318, Ser. I, 157–160.
Malyshev V. and Spieksma F.M. (1997): Dynamics in neural networks with a finite number of patterns, to appear in Math. Phys. Electr. J.
Newman C. (1988): Memory capacity in neural network models. Rigorous bounds, Neural Networks 1, 223–238.
Peretto P. (1992): An introduction to the modeling of neural networks, Cambridge University Press.
Pastur L. and Figotin A. (1977): Sov. J. Low Temp. Phys. 3, 378–383.
Pastur L. and Figotin A. (1978): Theor. Math. Phys. 35, 403–414.
Pastur L. and Shcherbina M. (1991): J. Stat. Phys. 621, 1–19.
Pastur L., Shcherbina M. and Tirozzi B. (1994): J. Stat Phys. 74, 1161–1183.
Pastur L., Shcherbina M. and Tirozzi B. (1996): Disordered systems: self averaging and replica solutions, WIAS Workshop on “Mathematics of random interactions”, Berlin, 21–24 August 1996.
Procesi C. and Tirozzi B. (1989): Int. J. Mod. Phys., 134, 143–159.
Ritter H., Martinez Th. and Schulten K. (1994): Neurale Netze, Addeson, Bonn.
Rojas R. (1993): Theorie der neuronalen Netze, Springer.
Rosenblatt F. (1980): Principles of Neurodynamics, Spartina, New York.
Scacciatelli E. and Tirozzi B. (1992): J. Stat. Phys. 67, 981–1008.
Shcherbina M. (1997): On the replica symmetric solution for the Sherrington-Kirkpatrick model, to appear in Heiv. Phys. Acta.
Shcherbina M. and Tirozzi B. (1993): J. Stat. Phys. 72, 113–125.
Simon B. (1993): The statistical mechanics of lattice gases, Princeton University.
Sinai Ya. G. (1982): Model Theory of Phase transitions: Rigorous results, Oxford.
Sinai Ya. G. (1994): Topics in ergodic theory, Princeton University Press.
Tirozzi B. (1995): Modelli Matematici di Reti Neurali, CEDAM, Padova.
Talagrand M. (1996): The Sherrington-Kirkpatrick model: A challenge for mathematicians, Paris, Preprint.
Talagrand M. (1996): Rigorous results for the Hopfield model with many patterns, Paris VI, Preprint.
Taylor J.G. ed. (1993): Mathematical approaches to neural networks, North Holland, Amsterdam.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Albeverio, S., Tirozzi, B. (1997). An Introduction to the Mathematical Theory of Neural Networks. In: Garrido, P.L., Marro, J. (eds) Fourth Granada Lectures in Computational Physics. Lecture Notes in Physics, vol 493. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-14148-9_5
Download citation
DOI: https://doi.org/10.1007/978-3-662-14148-9_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-662-14150-2
Online ISBN: 978-3-662-14148-9
eBook Packages: Springer Book Archive