Multiple photons enhance quantum machine learning

Многократные фотоны усиливают квантовое машинное обучение
Roberto Osellame, Tobias Haug, Simone Piacentini, Andrea Crespi, Francesco Ceccarelli, Yong Wang, Zhenghao Yin, Ciro Pentangelo, Philip Walther
2026-06-19

linear optical circuitsmetric learningmulti-photon statesphotonic quantum machine learningquantum Fisher information
Abstract Photons are promising candidates for quantum information technology due to their high robustness and long coherence time at room temperature. Inspired by recent advances in photonic computing techniques, research has increasingly turned to quantum machine learning (QML) on photonic platforms. Although photons provide a high-dimensional quantum feature space suitable for computation, a general understanding of how photon number can be harnessed for learning tasks remains limited. Here, we establish both theoretically and experimentally a learning-capacity advantage of multi-photon states over single-photon states in photonic QML. We prove that the learning capacity of linear optical circuits, quantified by the rank of the data quantum Fisher information matrix, scales polynomially with the photon number. This scaling enables multi-photon models to generalize from fewer training states than corresponding single-photon models and to achieve lower test loss under the same architecture and learning protocol. Moreover, we experimentally corroborate these findings through unitary learning and metric learning tasks, by performing online training on a fully programmable photonic integrated platform. Our work highlights the potential of photonic QML and paves the way for achieving quantum enhancement in practical machine learning applications.
1
Because of this polynomial scaling, multi-photon models generalize from fewer training states than equivalent single-photon models.
2
Experimental online training on a programmable photonic integrated platform corroborates the theoretical advantages in unitary learning and metric learning tasks.
3
Multi-photon states provide a provable learning-capacity advantage over single-photon states in photonic quantum machine learning.
4
The learning capacity of linear optical circuits, measured by the rank of the data quantum Fisher information matrix, scales polynomially with photon number.
5
Under the same architecture and learning protocol, multi-photon models achieve lower test loss than corresponding single-photon models.

Multi-photon states in photonic quantum machine learning implemented with linear optical circuits on a programmable photonic integrated platform

Learning-capacity advantage (scaling of learning capacity quantified by rank of the quantum Fisher information matrix with photon number) and consequent generalization and test-loss improvement of multi-photon models versus single-photon models in photonic QML, supported by theoretical proof and experimental unitary and metric learning tasks

Publication Details
Publication Date
2026-06-19
Journal
Publisher
ISSN
Cited by
0
Access Type
Author Information
Authors
Roberto Osellame
Tobias Haug
Simone Piacentini
Andrea Crespi
Francesco Ceccarelli
Yong Wang
Zhenghao Yin
Ciro Pentangelo
Philip Walther
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%