Please use this identifier to cite or link to this item:
http://hdl.handle.net/2445/195455
Title: | Supervised learning of few dirty bosons with variable particle number |
Author: | Mujal Torreblanca, Pere Martínez Miguel, Alex Polls Martí, Artur Juliá-Díaz, Bruno Pilati, Sebastiano |
Keywords: | Bosons Física de partícules Teoria quàntica Bosons Particle physics Quantum theory |
Issue Date: | 24-Mar-2021 |
Publisher: | SciPost Foundation |
Abstract: | We investigate the supervised machine learning of few interacting bosons in optical speckle disorder via artificial neural networks. The learning curve shows an approximately universal power-law scaling for different particle numbers and for different interaction strengths. We introduce a network architecture that can be trained and tested on heterogeneous datasets including different particle numbers. This network provides accurate predictions for all system sizes included in the training set and, by design, is suitable to attempt extrapolations to (computationally challenging) larger sizes. Notably, a novel transfer-learning strategy is implemented, whereby the learning of the larger systems is substantially accelerated and made consistently accurate by including in the training set many small-size instances. |
Note: | Reproducció del document publicat a: https://doi.org/10.21468/SciPostPhys.10.3.073 |
It is part of: | SciPost Physics, 2021, vol. 10, num. 3, p. 73-89 |
URI: | http://hdl.handle.net/2445/195455 |
Related resource: | https://doi.org/10.21468/SciPostPhys.10.3.073 |
ISSN: | 2542-4653 |
Appears in Collections: | Articles publicats en revistes (Institut de Ciències del Cosmos (ICCUB)) Articles publicats en revistes (Física Quàntica i Astrofísica) |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
711452.pdf | 912.14 kB | Adobe PDF | View/Open |
This item is licensed under a
Creative Commons License