MATH 7870-001 — Randomized NLA for Data Science and Machine Learning


Spring 2026


Instructor: Akil Narayan
Email: akil(-at-)sci.utah.edu
Office phone: +1 801-581-8984
Office location: WEB 4666, LCB 116
Office hours: Tuesday 9:45-10:40am in LCB 116, or by appointment


Class meeting time: Tuesday, Thursday 10:45am - 12:05pm
Class meeting location: M LI 1735


This course is a gentle journey through the land of randomized methods, focused on core (numerical) linear algebra tasks, and if time permits with some exploration of stochastic optimization methods.

The course syllabus is here: PDF



The content of this course is split across this website and Canvas. The material available on this website is:
  • Course syllabus
  • Homework assignments
  • Lecture slides
  • Miscellaneous handouts and resources
The material available on Canvas is:
  • Course syllabus
  • Homework assignments and submission portal
  • Grades



In-class presentations have the schedule below.
NOTE: There are 3 slots per day. All classes are in-person, except Tues April 7, which will be over Zoom. A link will be advertised via email and also posted as a Canvas announcement.
Date Presenter Resources Topic
Thursday, April 2 Zane C. Ref Trace estimation (XTrace)
Tory R. Ref Fast Randomized Iteration
Gaurav D.
Tuesday, April 7 Anwesa D. ZOOM Ref Iterative Hessian Sketch
Tim S. ZOOM Ref Fast direct methods for Gaussian processes
Laurel W. ZOOM Ref Sparse Randomized Kaczmarz for EEG Signals
ZOOM
Thursday, April 9 Lucy L. Ref Randomized Cholesky-QR factorizations
Asher M.
Justin S. Ref Stochastic matrix inversion
Tuesday, April 14 Erin S. Ref Matrix sketching for linear mixed models
Zijie L. Ref Sharp analysis of low-rank kernel matrix approximations
Valeria S. Ref 1, Ref 2 Sampling the Unseen: Importance Weighting for Latent Confounders
Thursday, April 16 Joel K. Ref Stochastic reformulations of linear systems
Parikshita G.
Nicole L.
David R.
Tuesday, April 21 Eli F. Ref Trace estimation (Hutch++)
Shridhar V. Ref Randomly pivoted Cholesky
John T. Ref Recursive Sampling for the Nyström methods



Graded assignments


Individual grades for each assignment will be posted to Canvas. (uNID login required.) Note that the letter grades appearing on Canvas are not representative of predicted final letter grades for the course. Final letter grades will be computed according to the rubric and policies on the syllabus.






Miscellaneous handouts


The following are slides from class.

Description Posting date Download Marked document
Presentation information and a non-comprehensive list of papers March 4, 2026 PDF
00: Course overview January 03, 2026 PDF
01: An intro to numerical linear algebra January 07, 2026 PDF PDF
02: A review of probability January 12, 2026 PDF PDF
03: Matrix multiplication: preasymptotic estimation January 19, 2026 PDF PDF
04: Scalar concentration January 26, 2026 PDF PDF
05: Applications of scalar concentration February 2, 2026 PDF PDF
06: Matrix concentration: Initial results February 13, 2026 PDF PDF
07: Matrix concentration: Matrix Chernoff February 25, 2026 PDF PDF
08: Matrix concentration: Matrix Bernstein March 2, 2026 PDF PDF
09: Random embeddings March 24, 2026 PDF PDF