ALoRA: Hardware-Aware Fine Tuning for Photonic Large Neural Networks
We investigated the impact of analog errors on large-scale photonic neural networks using an experimentally obtained error model. Performance degradation due to analog error can be recovered by proposed hardware-aware fine-tuning with ∼1.6% trainable parameters.
Authors
- M. Nakajima (ORCID: https://orcid.org/0000-0001-9451-4675)
- T. Taniguchi (ORCID: https://orcid.org/0000-0001-9865-1351)
- Satoshi Kawakami (ORCID: https://orcid.org/0000-0001-5044-744X)
- Toshikazu Hashimoto (ORCID: https://orcid.org/0000-0002-1661-2736)
- Kohei Ikeda (ORCID: https://orcid.org/0009-0001-0689-934X)
Institutions
- Kyushu University (JP)
- NTT (Japan) (JP)
Publication Details
- Published
- 2025-06-29
- DOI
- https://doi.org/10.23919/oecc/psc62146.2025.11109873
- Primary Topic
- Neural Networks and Reservoir Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00