Trainability barriers and opportunities in quantum generative modeling
npj Quantum Information · 13.11.2024 · 10 (1)
DOI: 10.1038/s41534-024-00902-0 ↗Kurz gefasst
Fachartikel in der Zeitschrift npj Quantum Information, online erschienen am 13.11.2024. Verfasst von Manuel S. Rudolph und 7 weiteren Autor:innen. Die Arbeit wurde bisher 47-mal zitiert (Crossref) und nennt 114 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Sandoz Family Foundation-Monique de Meuron program for Academic Promotion., CERN.
Abstract (Originalsprache)
Quantum generative models provide inherently efficient sampling strategies and thus show promise for achieving an advantage using quantum hardware. In this work, we investigate the barriers to the trainability of quantum generative models posed by barren plateaus and exponential loss concentration. We explore the interplay between explicit and implicit models and losses, and show that using quantum generative models with explicit losses such as the KL divergence leads to a new flavor of barren plateaus. In contrast, the implicit Maximum Mean Discrepancy loss can be viewed as the expectation value of an observable that is either low-bodied and provably trainable, or global and untrainable depending on the choice of kernel. In parallel, we find that solely low-bodied implicit losses cannot in general distinguish high-order correlations in the target data, while some quantum loss estimation strategies can. We validate our findings by comparing different loss functions for modeling data from High-Energy-Physics.
Förderung
- Sandoz Family Foundation-Monique de Meuron program for Academic Promotion.
- CERN
Zitieren
APA
Rudolph, M. S., Lerch, S., Thanasilp, S., Kiss, O., Shaya, O., Vallecorsa, S., Grossi, M., & Holmes, Z. (2024). Trainability barriers and opportunities in quantum generative modeling. npj Quantum Information, 10(1). https://doi.org/10.1038/s41534-024-00902-0
Vancouver
Rudolph MS, Lerch S, Thanasilp S, Kiss O, Shaya O, Vallecorsa S, et al. Trainability barriers and opportunities in quantum generative modeling. npj Quantum Information. 2024;10(1). doi:10.1038/s41534-024-00902-0
BibTeX
@article{Rudolph2024,
author = {Rudolph, Manuel S. and Lerch, Sacha and Thanasilp, Supanut and Kiss, Oriel and Shaya, Oxana and Vallecorsa, Sofia and Grossi, Michele and Holmes, Zoë},
title = {Trainability barriers and opportunities in quantum generative modeling},
journal = {npj Quantum Information},
year = {2024},
volume = {10},
number = {1},
doi = {10.1038/s41534-024-00902-0}
}
EUVIMED