PCA meets RG

Abstract

A system with many degrees of freedom can be characterized by a covariance matrix; principal components analysis (PCA) focuses on the eigenvalues of this matrix, hoping to find a lower dimensional description. But when the spectrum is nearly continuous, any distinction between components that we keep and those that we ignore becomes arbitrary; it then is natural to ask what happens as we vary this arbitrary cutoff. We argue that this problem is analogous to the momentum shell renormalization group (RG). Following this analogy, we can define relevant and irrelevant operators, where the role of dimensionality is played by properties of the eigenvalue density. These results also suggest an approach to the analysis of real data. As an example, we study neural activity in the vertebrate retina as it responds to naturalistic movies, and find evidence of behavior controlled by a nontrivial fixed point. Applied to financial data, our analysis separates modes dominated by sampling noise from a smaller but still macroscopic number of modes described by a non--Gaussian distribution.

8 pages, 4 figures

References in corpus (8)

• A Tutorial on Principal Component Analysis
• More is the Same; Phase Transitions and Mean Field Theories
• Cleaning large correlation matrices: tools from random matrix theory
• The sloppy model universality class and the Vandermonde matrix
• Rotational invariant estimator for general noisy matrices
• Spin Glass in a Field: a New Zero-Temperature Fixed Point in Finite Dimensions
• Reflections on Gibbs: From Statistical Physics to the Amistad
• Innovations in Statistical Physics

Cited by in corpus (38)

• Machine learning and the physical sciences
• A high-bias, low-variance introduction to Machine Learning for physicists
• Discovering Phases, Phase Transitions and Crossovers through Unsupervised Machine Learning: A critical examination
• Neural Network Renormalization Group
• Coarse--graining, fixed points, and scaling in a large population of neurons
• Learning Relevant Features of Data with Multi-scale Tensor Networks
• Dynamical renormalization group approach to the collective behaviour of swarms
• Nonperturbative renormalization for the neural network-QFT correspondence
• Deep learning black hole metrics from shear viscosity
• The higher-order spectrum of simplicial complexes: a renormalization group approach
• Theoretical ecology without species
• Critical phenomena in the temperature-pressure-crowding phase diagram of a protein
• Field theoretical approach for signal detection in nearly continuous positive spectra II: Tensorial data
• Renormalization group crossover in the critical dynamics of field theories with mode coupling terms
• Signal detection in nearly continuous spectra and symmetry breaking
• Field theoretical approach for signal detection in nearly continuous positive spectra I: Matricial data
• Inferring relevant features: from QFT to PCA
• An analysis of localization transitions using non-parametric unsupervised learning
• Examples of renormalization group transformations for image sets
• On Koopman Mode Decomposition and Tensor Component Analysis
• Renormalization Group as a Koopman Operator
• Coarse--graining and hints of scaling in a population of 1000+ neurons
• Optimal encoding in stochastic latent-variable Models
• The statistical mechanics of Twitter
• Criticality at work: scaling in the mouse cortex enhances performance
• Large time effective kinetics $\beta$-functions for quantum (2+p)-spin glass
• Correlations in the shear flow of athermal amorphous solids: A principal component analysis
• Lateral predictive coding revisited: Internal model, symmetry breaking, and response time
• Constructive expansion for quartic vector fields theories. I. Low dimensions
• Functional renormalization group for signal detection and stochastic ergodicity breaking
• Wilsonian Renormalization of Neural Network Gaussian Processes
• Phase transitions in in vivo or in vitro populations of spiking neurons belong to different universality classes
• Functional Renormalization Group Approach for Signal Detection
• An intriguing connection between Pisarski's fixed point and (2+3)-spin glasses
• Signal inference in financial stock return correlations through phase-ordering kinetics in the quenched regime
• Functional Renormalization for Signal Detection: Dimensional Analysis and Dimensional Phase Transition for Nearly Continuous Spectra Effective Field Theory
• Gaussian Information Bottleneck and the Non-Perturbative Renormalization Group
• Renormalization group for spectral collapse in random matrices with power-law variance profiles