Current browse context:
quant-ph
Change to browse by:
References & Citations
Quantum Physics
Title: Manifold learning via quantum dynamics
(Submitted on 20 Dec 2021 (v1), last revised 12 Jan 2022 (this version, v2))
Abstract: We introduce an algorithm for computing geodesics on sampled manifolds that relies on simulation of quantum dynamics on a graph embedding of the sampled data. Our approach exploits classic results in semiclassical analysis and the quantum-classical correspondence, and forms a basis for techniques to learn the manifold from which a dataset is sampled, and subsequently for nonlinear dimensionality reduction of high-dimensional datasets. We illustrate the new algorithm with data sampled from model manifolds and also by a clustering demonstration based on COVID-19 mobility data. Finally, our method reveals interesting connections between the discretization provided by data sampling and quantization.
Submission history
From: Mohan Sarovar [view email][v1] Mon, 20 Dec 2021 18:54:03 GMT (7716kb,D)
[v2] Wed, 12 Jan 2022 06:18:42 GMT (7716kb,D)
Link back to: arXiv, form interface, contact.