AMS :: Fall Western Sectional Meeting (formerly at University of New Mexico), AMS Special Session

AMS Sectional Meeting AMS Special Session

Current as of Saturday, October 30, 2021 03:30:04

Inquiries: meet@ams.org

Fall Western Sectional Meeting (formerly at University of New Mexico)

now meeting virtually, PDT (hosted by the American Mathematical Society), Virtual, RI
October 23-24, 2021 (Saturday - Sunday)
Meeting #1172

Associate Secretary for the AMS Scientific Program:

Michel L Lapidus, AMS lapidus@math.ucr.edu

Special Session on Theoretical and Applied perspectives in Machine Learning

Saturday October 23, 2021, 8:00 a.m.-11:50 a.m.
Special Session on Theoretical and Applied perspectives in Machine Learning, I
Special Session 17, American Mathematical Society
Organizers:
Jehanzeb Hameed Chaudhary, University of New Mexico
Adam Thomas Rupe, Los Alamos National Laboratory
Simon Tavener, Colorado State University tavener@math.colostate.edu
Dimiter Vassilev, University of New Mexico
Velimir Vesselinov, Los Alamos National Laboratory
- 8:00 a.m.
  Finding symmetry breaking order parameters with Euclidean neural networks.
  Tess Smidt*, Massachusetts Institute of Technology
  Mario Geiger, École polytechnique fédérale de Lausanne
  Benjamin K. Miller, University of Amsterdam
  (1172-20-278)
- 8:30 a.m.
  Frank-Wolfe methods for efficient geodesically convex optimization.
  Melanie Weber*, Princeton University
  Suvrit Sra, Massachusetts Institute of Technology
  (1172-49-138)
- 9:00 a.m.
  Geometry inspired DNNs on Manifold-structured Data.
  Rongjie Lai*, Rensselaer Polytechnic Institute
  (1172-68-176)
- 9:30 a.m.
  Break.
- 10:00 a.m.
  Reinforcement learning and string theory geometry.
  Andre Lukas*, University of Oxford
  (1172-58-256)
- 10:30 a.m.
  Learning vector fields and diffferential forms with the spectral exterior calculus.
  Dimitrios Giannakis*, Dartmouth College
  (1172-58-322)
- 11:00 a.m.
  Using generative adversarial networks to produce knots with specified invariants.
  Amy Eubanks, Brigham Young University
  Mark Hughes*, Brigham Young University
  Jared Slone, Brigham Young University
  (1172-54-13)
- 11:30 a.m.
  The approximation theory of shallow neural networks.
  Jonathan W. Siegel*, The Pennsylvania State University
  (1172-62-356)

Saturday October 23, 2021, 2:00 p.m.-6:00 p.m.
Special Session on Theoretical and Applied perspectives in Machine Learning, II
Special Session 17, American Mathematical Society
Organizers:
Jehanzeb Hameed Chaudhary, University of New Mexico
Adam Thomas Rupe, Los Alamos National Laboratory
Simon Tavener, Colorado State University tavener@math.colostate.edu
Dimiter Vassilev, University of New Mexico
Velimir Vesselinov, Los Alamos National Laboratory
- 2:00 p.m.
  Efficient training of infinite-depth neural networks via Jacobian-Free backpropagation.
  Samy Wu Fung*, Colorado School of Mines
  Howard Heaton, University of California, Los Angeles
  Qiuwei Li, University of California, Los Angeles
  Daniel McKenzie, University of California, Los Angeles
  Stanley Osher, University of California, Los Angeles
  Wotao Yin, Alibaba
  (1172-65-164)
- 2:30 p.m.
  Convolutional neural network for solving inverse problems of nonlinear wave equations.
  Gunther Uhlmann, University of Washington
  Yiran Wang*, Emory University
  (1172-35-187)
- 3:00 p.m.
  Optimal design of inverse problems governed by PDEs.
  Alen Alexanderian*, North Carolina State University
  (1172-65-293)
- 3:30 p.m.
  Break.
- 4:00 p.m.
  Parallel-in-time training of recurrent neural networks.
  Eric C. Cyr*, Sandia National Laboratories
  Gordon E. Moon, Korea Aerospace University
  (1172-65-330)
- 4:30 p.m.
  Neural operator: Learning maps between function spaces.
  Zongyi Li*, Caltech
  (1172-65-359)
- 5:00 p.m.
  Hidden variables and inference for linear non-Gaussian causal models.
  Elina Robeva*, University of British Columbia
  (1172-62-353)
- 5:30 p.m.
  Discussion.

Inquiries: meet@ams.org