On the criticality of inferred models

Abstract

Advanced inference techniques allow one to reconstruct the pattern of interaction from high dimensional data sets. We focus here on the statistical properties of inferred models and argue that inference procedures are likely to yield models which are close to a phase transition. On one side, we show that the reparameterization invariant metrics in the space of probability distributions of these models (the Fisher Information) is directly related to the model's susceptibility. As a result, distinguishable models tend to accumulate close to critical points, where the susceptibility diverges in infinite systems. On the other, this region is the one where the estimate of inferred parameters is most stable. In order to illustrate these points, we discuss inference of interacting point processes with application to financial data and show that sensible choices of observation time-scales naturally yield models which are close to criticality.

6 pages, 2 figures, version to appear in JSTAT

References in corpus (8)

• Identification of direct residue contacts in protein-protein interaction by message passing
• Market impact and trading profile of large trading orders in stock markets
• The Ising Model for Neural Data: Model Quality and Approximate Methods for Extracting Functional Connectivity
• Statistical physics of pairwise probability models
• Small-correlation expansions for the inverse Ising problem
• Specialization of strategies and herding behavior of trading firms in a financial market
• Inference algorithms for gene networks: a statistical mechanics analysis
• Turnover, account value and diversification of real traders: evidence of collective portfolio optimizing behavior

Cited by in corpus (41)

• Colloquium: Criticality and dynamical scaling in living systems
• Inverse statistical problems: from the inverse Ising problem to data science
• Information-based fitness and the emergence of criticality in living systems
• Finite-size scaling as a way to probe near-criticality in natural swarms
• Zipf's law and criticality in multivariate data without fine-tuning
• Quantum-Assisted Learning of Hardware-Embedded Probabilistic Graphical Models
• Self-organization toward criticality by synaptic plasticity
• On sampling and modeling complex systems
• Determination of the edge of criticality in echo state networks through Fisher information maximization
• Thermodynamics of Restricted Boltzmann Machines and related learning dynamics
• Signatures of criticality arise in simple neural population models with correlations
• Statistical Criticality arises in Most Informative Representations
• Maximum entropy models reveal the excitatory and inhibitory correlation structures in cortical neuronal activity
• Maximum likelihood reconstruction for Ising models with asynchronous updates
• Cumulants of Hawkes point processes
• Consistent inference of a general model using the pseudo-likelihood method
• Quantifying Relevance in Learning and Inference
• Entropy evaluation sheds light on ecosystem complexity
• Balancing selfishness and norm conformity can explain human behavior in large-scale Prisoner's Dilemma games and can poise human groups near criticality
• Detecting intraday financial market states using temporal clustering
• Ensemble Inhibition and Excitation in the Human Cortex: an Ising Model Analysis with Uncertainties
• A statistical physics perspective on criticality in financial markets
• Adaptive Thouless-Anderson-Palmer approach to inverse Ising problems with quenched random fields
• Cooperation, competition and the emergence of criticality in communities of adaptive systems
• Heterogeneity extends criticality
• Dynamical signatures of structural connectivity damage to a model of the brain posed at criticality
• Spatial patterns emerging from a stochastic process near criticality
• Financial interaction networks inferred from traded volumes
• Inferring couplings in networks across order-disorder phase transitions
• Large Deviations Properties of Maximum Entropy Markov Chains from Spike Trains
• The free lunch of a scale-free metabolism
• Biases in Inverse Ising Estimates of Near-Critical Behaviour
• Learning performance in inverse Ising problems with sparse teacher couplings
• Optimal encoding in stochastic latent-variable Models
• Unlearning regularization for Boltzmann Machines
• The statistical mechanics of Twitter
• Dataset-Free Weight-Initialization on Restricted Boltzmann Machine
• Thermodynamic Formalism in Neuronal Dynamics and Spike Train Statistics
• The dual of the space of interactions in neural network models
• Defining and finding lifelike entities with a lazy filter
• Agglomerative Likelihood Clustering