Differentiable Neural Architecture Search for Photonic-Aware Neural Networks
We propose a differentiable Neural Architecture Search framework for photonic-aware neural networks integrating photonic constraints and quantization-aware training. The resulting architectures outperform handcrafted models achieving 79% accuracy at 6-bit quantization, closely matching full-precision baselines.
Authors
- Nicola Andriolli (ORCID: https://orcid.org/0000-0001-8322-6854)
- Emilio Paolini (ORCID: https://orcid.org/0000-0001-7486-9268)
- Marco Cococcioni (ORCID: https://orcid.org/0000-0002-7020-1524)
- Gianluca Paolini (ORCID: https://orcid.org/0000-0002-8266-362X)
Institutions
- University of Pisa (IT)
- Scuola Superiore Sant'Anna (IT)
Publication Details
- Published
- 2025-06-29
- DOI
- https://doi.org/10.23919/oecc/psc62146.2025.11109505
- Primary Topic
- Neural Networks and Reservoir Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00