An Almost Sure Large Deviation Principle for the Hopfield Model
نویسندگان
چکیده
We prove a large deviation principle for the finite dimensional marginals of the Gibbs distribution of the macroscopic ‘overlap’-parameters in the Hopfield model in the case where the number of random ‘patterns’, M , as a function of the system size N satisfies lim supM(N)/N = 0. In this case the rate function is independent of the disorder for almost all realizations of the patterns.
منابع مشابه
Rigorous Results on the Hopfield Model Ofneural Networks
We review some recent rigorous results in the theory of neural networks, and in particular on the thermodynamic properties of the Hoppeld model. In this context, the model is treated as a Curie-Weiss model with random interactions and large deviation techniques are applied. The tractability of the random interactions depends strongly on how the number, M, of stored patterns scales with the size...
متن کاملThe storage capacity of the Hopfield model and moderate deviation principles
This note relates the storage capacity of the Hopfield model of neural networks to the existence of a moderate deviation principle for the empirical correlation of the patterns. This moderate deviation principle is satisfied under a certain condition on the moment generating function of these correlations which on the other hand can be verified in many cases by GHSand FKG-type inequalities. Exa...
متن کاملSelf-averaging in a Class of Generalized Hopfield Models#
We prove the almost sure convergence to zero of the uctuations of the free energy, resp. local free energies, in a class of disordered mean-eld spin systems that generalize the Hoppeld model in two ways: 1) Multi-spin interactions are permitted and 2) the random variables i describing thèpatterns' can have arbitrary distributions with mean zero and nite 4+-th moments. The number of patterns, M,...
متن کاملAn Almost Sure Central Limit Theorem for the Hopfield Model#
We prove a central limit theorem for the nite dimensional marginals of the Gibbs distribution of the macroscopicòverlap'-parameters in the Hoppeld model in the case where the number of random`patterns', M, as a function of the system size N satisses lim N"1 M(N)=N = 0, without any assumptions on the speed of convergence. The covariance matrix of the limiting gaussian distributions is diagonal a...
متن کاملRigorous Solution of a Mean Field Spin Glass Model
A separable spin glass model whose exchange integral takes the form Jij-J(ilj2+12jl) which was solved by van Hemmen et al. [12] using large deviation theory [14] is rigorously treated. The almost sure convergence criteria associated with the cumulant generating function C(t) with respect to the quenched random variables (is carefully investigated, and it is proved that the related excluded null...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 1995