Details
| Original language | English |
|---|---|
| Pages (from-to) | 2779-2786 |
| Number of pages | 8 |
| Journal | Nanophotonics |
| Volume | 14 |
| Issue number | 16 |
| Early online date | 23 Jun 2025 |
| Publication status | Published - 2 Aug 2025 |
Abstract
Optical artificial neural networks (OANNs) leverage the advantages of photonic technologies including high processing speeds, low energy consumption, and mass production to establish a competitive and scalable platform for machine learning applications. While recent advancements have focused on harnessing spatial or temporal modes of light, the frequency domain attracts a lot of attention, with current implementations including spectral multiplexing, neural networks in nonlinear optical systems and extreme learning machines. Here, we present an experimental realization of a programmable photonic frequency circuit, realized with fiber-optical components, and implement the in-situ training with optical weight control of an OANN operating in the frequency domain. Input data is encoded into phases of frequency comb modes, and programmable phase and amplitude manipulations of the spectral modes enable in-situ training of the OANN, without employing a digital model of the device. The trained OANN achieves multiclass classification accuracies exceeding 90 %, comparable to conventional machine learning approaches. This proof-of-concept demonstrates the feasibility of a multilayer OANN in the frequency domain and can be extended to a scalable, integrated photonic platform with ultrafast weights updates, with potential applications to single-shot classification in spectroscopy.
Keywords
- machine learning, photonic computing, ultrafast optics
ASJC Scopus subject areas
- Biochemistry, Genetics and Molecular Biology(all)
- Biotechnology
- Materials Science(all)
- Electronic, Optical and Magnetic Materials
- Physics and Astronomy(all)
- Atomic and Molecular Physics, and Optics
- Engineering(all)
- Electrical and Electronic Engineering
Sustainable Development Goals
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Nanophotonics, Vol. 14, No. 16, 02.08.2025, p. 2779-2786.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - In-situ training in programmable photonic frequency circuits
AU - Rübeling, Philip
AU - Marchukov, Oleksandr V.
AU - Bellotti, Filipe F.
AU - Hoff, Ulrich B.
AU - Zinner, Nikolaj T.
AU - Kues, Michael
N1 - Publisher Copyright: © 2025 the author(s), published by De Gruyter, Berlin/Boston 2025.
PY - 2025/8/2
Y1 - 2025/8/2
N2 - Optical artificial neural networks (OANNs) leverage the advantages of photonic technologies including high processing speeds, low energy consumption, and mass production to establish a competitive and scalable platform for machine learning applications. While recent advancements have focused on harnessing spatial or temporal modes of light, the frequency domain attracts a lot of attention, with current implementations including spectral multiplexing, neural networks in nonlinear optical systems and extreme learning machines. Here, we present an experimental realization of a programmable photonic frequency circuit, realized with fiber-optical components, and implement the in-situ training with optical weight control of an OANN operating in the frequency domain. Input data is encoded into phases of frequency comb modes, and programmable phase and amplitude manipulations of the spectral modes enable in-situ training of the OANN, without employing a digital model of the device. The trained OANN achieves multiclass classification accuracies exceeding 90 %, comparable to conventional machine learning approaches. This proof-of-concept demonstrates the feasibility of a multilayer OANN in the frequency domain and can be extended to a scalable, integrated photonic platform with ultrafast weights updates, with potential applications to single-shot classification in spectroscopy.
AB - Optical artificial neural networks (OANNs) leverage the advantages of photonic technologies including high processing speeds, low energy consumption, and mass production to establish a competitive and scalable platform for machine learning applications. While recent advancements have focused on harnessing spatial or temporal modes of light, the frequency domain attracts a lot of attention, with current implementations including spectral multiplexing, neural networks in nonlinear optical systems and extreme learning machines. Here, we present an experimental realization of a programmable photonic frequency circuit, realized with fiber-optical components, and implement the in-situ training with optical weight control of an OANN operating in the frequency domain. Input data is encoded into phases of frequency comb modes, and programmable phase and amplitude manipulations of the spectral modes enable in-situ training of the OANN, without employing a digital model of the device. The trained OANN achieves multiclass classification accuracies exceeding 90 %, comparable to conventional machine learning approaches. This proof-of-concept demonstrates the feasibility of a multilayer OANN in the frequency domain and can be extended to a scalable, integrated photonic platform with ultrafast weights updates, with potential applications to single-shot classification in spectroscopy.
KW - machine learning
KW - photonic computing
KW - ultrafast optics
UR - http://www.scopus.com/inward/record.url?scp=105009132590&partnerID=8YFLogxK
U2 - 10.1515/nanoph-2025-0125
DO - 10.1515/nanoph-2025-0125
M3 - Article
AN - SCOPUS:105009132590
VL - 14
SP - 2779
EP - 2786
JO - Nanophotonics
JF - Nanophotonics
SN - 2192-8606
IS - 16
ER -