Learning emergent models from ab initio many-body calculations.
Title: Learning emergent models from ab initio many-body calculations.
Speaker: Prof. Lucas K. Wagner (The Grainger College of Engineering, University of Illinois Urbana-Champaign)
Start Date/Time: 2023-04-26 / 09:00 (Taipei time)
End Date/Time: 2023-04-26 / 10:00
Host: Prof. Chia-Min Chung (NSYSU)
Online Zoom Link: https://us02web.zoom.us/j/86867205231?pwd=OTJVTURuVU9FVzkzR01kMVUwcGVvZz09
[Registration] is required
A crucial step for understanding emergent physics of correlated systems is determining whether a particular model is emergent from ultraviolet physics. So far, a systematic way of detecting emergent behavior from ab initio calculations has yet to be available in the state-of-the-art model derivation procedures. In this study, we propose that low-energy emergent physics forms disjoint clusters in an extended descriptor space beyond just energy cutoffs. Working from highly-accurate many-body quantum Monte Carlo calculations of the ab initio hydrogen chains, we show that a cluster-based classification algorithm can detect a Heisenberg-like emergent subspace using a set of descriptors of the energy eigenstates.