Bulletin of the American Physical Society
APS March Meeting 2022
Volume 67, Number 3
Monday–Friday, March 14–18, 2022; Chicago
Session G14: Learning in Physical SystemsInvited Session Live Streamed
|
Hide Abstracts |
Sponsoring Units: DSOFT GSNP Chair: Menachem Stern, University of Pennsylvania Room: McCormick Place W-183B |
Tuesday, March 15, 2022 11:30AM - 12:06PM |
G14.00001: How a time-reversal-invariant physical system can be turned into a self-learning machine Invited Speaker: Florian Marquardt Machine learning using artifical neural networks is revolutionizing many areas of science and technology. This increases the urgency for exploring alternatives to artificial neural networks running on digital hardware. These alternatives might eventually be faster and/or more power-efficient. With this in mind, we ask the question whether one can identify a general principle that would enable a nonlinear physical system to become a self-learning machine - i.e. a physical information-processing device where internal degrees of freedom self-adjust by physical interactions to learn a desired input-output relation. In this talk, I will present our recent idea on how this might be achieved for arbitrary time-reversal-invariant Hamiltonian systems. I will introduce the principle of 'Hamiltonian Echo Backpropagation', and demonstrate how efficient learning could be possible in a wide class of physical systems. |
Tuesday, March 15, 2022 12:06PM - 12:42PM |
G14.00002: Physics for neuromorphic computing Invited Speaker: Julie Grollier
|
Tuesday, March 15, 2022 12:42PM - 1:18PM |
G14.00003: Materials that learn from examples Invited Speaker: Arvind Murugan Learning is usually associated with neural networks. But non-neural systems can also accumulate incremental changes over time and thus respond better to future environments. We show how seemingly `dumb' physical systems like DNA crystals and elastic materials can learn to recognize complex patterns in chemical or mechanical stimuli, much like a neural network. We outline the potential and limits of such `mechanical intelligence' due to physically realizable learning dynamics. |
Tuesday, March 15, 2022 1:18PM - 1:54PM |
G14.00004: Multifunctional networks using local training rules Invited Speaker: Nidhi Pashine The mechanical properties of disordered networks can be significantly modified to exhibit unconventional responses by changing a small fraction of their bonds. One such response is a long distance 'allosteric' response where an applied local strain at one site of the system creates a localized output strain at a distant site. Previous work has relied on computer simulations to design and predict the response of such systems using a cost function where the response of the entire network to each bond removal is used at each step to determine which bond to prune. I will present a novel design approach that relies only on local stress measurements to incorporate allosteric responses. Instead of relying on computer simulations, we have an experimental method to measure local stress distributions using photoelastic measurements. This approach can be used to experimentally implement different pruning methods. In order to create an allosteric response, instead of completely removing a set of bonds, one can also soften a fraction of the network's bonds. By creating some bonds out of variable stiffness composites that can readily switch between stiff and soft states, we can activate a particular allosteric response in a system. We can also create allosteric networks that exhibit different functionalities by softening different sets of bonds. This work provides a path to understand and create adaptable and trainable allosteric metamaterials. |
Tuesday, March 15, 2022 1:54PM - 2:30PM |
G14.00005: Decentralized, Physics-Driven Learning Invited Speaker: Sam J Dillavou
|
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2025 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700