-
Central limit theorem for the overlaps on the Nishimori line
Authors:
Francesco Camilli,
Pierluigi Contucci,
Emanuele Mingione
Abstract:
The overlap distribution of the Sherrington-Kirkpatrick model on the Nishimori line has been proved to be self averaging for large volumes. Here we study the joint distribution of the rescaled overlaps around their common mean and prove that it converges to a Gaussian vector.
The overlap distribution of the Sherrington-Kirkpatrick model on the Nishimori line has been proved to be self averaging for large volumes. Here we study the joint distribution of the rescaled overlaps around their common mean and prove that it converges to a Gaussian vector.
△ Less
Submitted 25 June, 2023; v1 submitted 31 May, 2023;
originally announced May 2023.
-
Human-AI ecosystem with abrupt changes as a function of the composition
Authors:
Pierluigi Contucci,
János Kertész,
Godwin Osabutey
Abstract:
The progressive advent of artificial intelligence machines may represent both an opportunity or a threat. In order to have an idea of what is coming we propose a model that simulate a Human-AI ecosystem. In particular we consider systems where agents present biases, peer-to-peer interactions and also three body interactions that are crucial and describe two humans interacting with an artificial ag…
▽ More
The progressive advent of artificial intelligence machines may represent both an opportunity or a threat. In order to have an idea of what is coming we propose a model that simulate a Human-AI ecosystem. In particular we consider systems where agents present biases, peer-to-peer interactions and also three body interactions that are crucial and describe two humans interacting with an artificial agent and two artificial intelligence agents interacting with a human. We focus our analysis by exploring how the relative fraction of artificial intelligence agents affect that ecosystem. We find evidence that for suitable values of the interaction parameters, arbitrarily small changes in such percentage may trigger dramatic changes for the system that can be either in one of the two polarised states or in an undecided state.
△ Less
Submitted 7 April, 2022;
originally announced April 2022.
-
Alienation in Italian cities. Social network fragmentation from collective data
Authors:
Pierluigi Contucci,
Cecilia Vernia
Abstract:
We study the structure of a social network of strong ties (trust network) investigating its property of connectedness versus fragmentation. To this purpose we analyse an extensive set of census data, about marrying or having children with immigrants, collected by Italian national statistical institute for all Italian municipalities from 2001 to 2011. Not using neither obtaining personal local info…
▽ More
We study the structure of a social network of strong ties (trust network) investigating its property of connectedness versus fragmentation. To this purpose we analyse an extensive set of census data, about marrying or having children with immigrants, collected by Italian national statistical institute for all Italian municipalities from 2001 to 2011. Not using neither obtaining personal local information but only average ones, our method fully complies with privacy and confidentiality. Our findings show that large cities display the behaviour of highly fragmented trust networks where individuals face possible phenomena of alienation. Smaller cities and villages instead behave like fully connected social systems with a rich tie structure, where isolation is rare or completely absent. While confirming classical sociological theories on alienation in large urban areas our approach provides a quantitative method to test them and a predictive tool for policy makers.
△ Less
Submitted 2 October, 2014;
originally announced October 2014.
-
On the equivalence of Hopfield Networks and Boltzmann Machines
Authors:
Adriano Barra,
Alberto Bernacchia,
Enrica Santucci,
Pierluigi Contucci
Abstract:
A specific type of neural network, the Restricted Boltzmann Machine (RBM), is implemented for classification and feature detection in machine learning. RBM is characterized by separate layers of visible and hidden units, which are able to learn efficiently a generative model of the observed data. We study a "hybrid" version of RBM's, in which hidden units are analog and visible units are binary, a…
▽ More
A specific type of neural network, the Restricted Boltzmann Machine (RBM), is implemented for classification and feature detection in machine learning. RBM is characterized by separate layers of visible and hidden units, which are able to learn efficiently a generative model of the observed data. We study a "hybrid" version of RBM's, in which hidden units are analog and visible units are binary, and we show that thermodynamics of visible units are equivalent to those of a Hopfield network, in which the N visible units are the neurons and the P hidden units are the learned patterns. We apply the method of stochastic stability to derive the thermodynamics of the model, by considering a formal extension of this technique to the case of multiple sets of stored patterns, which may act as a benchmark for the study of correlated sets. Our results imply that simulating the dynamics of a Hopfield network, requiring the update of N neurons and the storage of N(N-1)/2 synapses, can be accomplished by a hybrid Boltzmann Machine, requiring the update of N+P neurons but the storage of only NP synapses. In addition, the well known glass transition of the Hopfield network has a counterpart in the Boltzmann Machine: It corresponds to an optimum criterion for selecting the relative sizes of the hidden and visible layers, resolving the trade-off between flexibility and generality of the model. The low storage phase of the Hopfield model corresponds to few hidden units and hence a overly constrained RBM, while the spin-glass phase (too many hidden units) corresponds to unconstrained RBM prone to overfitting of the observed data.
△ Less
Submitted 10 January, 2012; v1 submitted 13 May, 2011;
originally announced May 2011.
