Data clustering based on Langevin annealing with a self-consistent potential

DOI10.1090/QAM/1521MaRDI QIDQ5383329zbMATH OpenOpenAlexWikidataFDO

Authors Kyle Lafata, Zhennan Zhou, Jian-Guo Liu, Fang-Fang Yin

Publication date 21 June 2019

Published in Quarterly of Applied Mathematics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1806.10597

zbMATH Keywords

Brownian motiondata clusteringergodicityGibbs distributionLangevin dynamicsquantum clusteringSchrödinger equation

Mathematics Subject Classification ID

Computational methods for ergodic theory (approximation of invariant measures, computation of Lyapunov exponents, entropy, etc.) (37M25) Stochastic ordinary differential equations (aspects of stochastic analysis) (60H10) Brownian motion (60J65) Classification and discrimination; cluster analysis (statistical aspects) (62H30) Closed and approximate solutions to the Schrödinger, Dirac, Klein-Gordon and other equations of quantum mechanics (81Q05) Stochastic methods (Fokker-Planck, Langevin, etc.) applied to problems in time-dependent statistical mechanics (82C31)

Abstract: This paper introduces a novel data clustering algorithm based on Langevin dynamics, where the associated potential is constructed directly from the data. To introduce a self-consistent potential, we adopt the potential model from the established Quantum Clustering method. The first step is to use a radial basis function to construct a density distribution from the data. A potential function is then constructed such that this density distribution is the ground state solution to the time-independent Schrodinger equation. The second step is to use this potential function with the Langevin dynamics at sub-critical temperature to avoid ergodicity. The Langevin equations take a classical Gibbs distribution as the invariant measure, where the peaks of the distribution coincide with minima of the potential surface. The time dynamics of individual data points lead to different metastable states, which are interpreted as cluster centers. Clustering is therefore achieved when subsets of the data aggregate - as a result of the Langevin dynamics for a moderate period of time - in the neighborhood of a particular potential minimum. While the data points are pushed towards potential minima by the potential gradient, Brownian motion allows them to effectively tunnel through local potential barriers and escape saddle points into locations of the potential surface otherwise forbidden. The algorithm's feasibility is first established based on several illustrating examples and theoretical analyses, followed by a stricter evaluation using a standard benchmark dataset.

Recommendations

• Data clustering with quantum mechanics
• A new approach for partitional clustering using entropy notation and Hopfield network
• Newtonian clustering: an approach based on molecular dynamics and global optimization
• Quantum annealing for combinatorial clustering
• Data clustering using hierarchical deterministic annealing and higher order statistics

Cites work

• Clustering of time series data -- a survey
• Clustering via Hilbert space
• Diffuse interface models on graphs for classification of high dimensional data
• Diffusion maps, spectral clustering and reaction coordinates of dynamical systems
• Exponentially accurate semiclassical dynamics: Propagation, localization, Ehrenfest times, scattering, and more general states.
• Geometric diffusions as a tool for harmonic analysis and structure definition of data: diffusion maps
• Homogenization limits and Wigner transforms
• scientific article; zbMATH DE number 2061791 (Why is no real title available?)
• scientific article; zbMATH DE number 1384314 (Why is no real title available?)
• scientific article; zbMATH DE number 835699 (Why is no real title available?)
• scientific article; zbMATH DE number 3261539 (Why is no real title available?)
• scientific article; zbMATH DE number 3280665 (Why is no real title available?)
• Hybrid deterministic-stochastic gradient Langevin dynamics for Bayesian learning
• Newtonian clustering: an approach based on molecular dynamics and global optimization
• Nonlinear nonequilibrium thermodynamics I. Linear and nonlinear fluctuation-dissipation theorems
• On Wigner measures
• Principal component analysis.
• Regularization on graphs with function-adapted diffusion processes

This page was built for publication: Data clustering based on Langevin annealing with a self-consistent potential

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5383329)