We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Quantum Physics

Title: Shining light on data: Geometric data analysis through quantum dynamics

Abstract: Experimental sciences have come to depend heavily on our ability to organize and interpret high-dimensional datasets. Natural laws, conservation principles, and inter-dependencies among observed variables yield geometric structure, with fewer degrees of freedom, on the dataset. We introduce the frameworks of semiclassical and microlocal analysis to data analysis and develop a novel, yet natural uncertainty principle for extracting fine-scale features of this geometric structure in data, crucially dependent on data-driven approximations to quantum mechanical processes underlying geometric optics. This leads to the first tractable algorithm for approximation of wave dynamics and geodesics on data manifolds with rigorous probabilistic convergence rates under the manifold hypothesis. We demonstrate our algorithm on real-world datasets, including an analysis of population mobility information during the COVID-19 pandemic to achieve four-fold improvement in dimensionality reduction over existing state-of-the-art and reveal anomalous behavior exhibited by less than 1.2% of the entire dataset. Our work initiates the study of data-driven quantum dynamics for analyzing datasets, and we outline several future directions for research.
Comments: Supplementary Material has high overlap with arXiv:2112.11161 by the same authors. v2 reorganizes presentation of results in paper
Subjects: Quantum Physics (quant-ph); Machine Learning (cs.LG); Statistics Theory (math.ST); Machine Learning (stat.ML)
Cite as: arXiv:2212.00682 [quant-ph]
  (or arXiv:2212.00682v2 [quant-ph] for this version)

Submission history

From: Mohan Sarovar [view email]
[v1] Thu, 1 Dec 2022 17:38:01 GMT (10656kb,D)
[v2] Tue, 13 Dec 2022 19:00:31 GMT (10660kb,D)

Link back to: arXiv, form interface, contact.