摘要
Deep learning has become a widespread tool in both science and industry. However, continued progress is hampered by the rapid growth in energy costs of ever-larger deep neural networks. Optical neural networks provide a potential means to solve the energy-cost problem faced by deep learning. Here, we experimentally demonstrate an optical neural network based on optical dot products that achieves 99% accuracy on handwritten-digit classification using similar to 3.1 detected photons per weight multiplication and similar to 90% accuracy using similar to 0.66 photons (similar to 2.5 x 10(-19) J of optical energy) per weight multiplication. The fundamental principle enabling our sub-photon-per-multiplication demonstration-noise reduction from the accumulation of scalar multiplications in dot-product sums-is applicable to many different optical-neural-network architectures. Our work shows that optical neural networks can achieve accurate results using extremely low optical energies.
学科与关键词
- Low-power scalable multilayer optoelectronic neural networks enabled with incoherent lightSong, Alexander*; Kottapalli, Sai Nikhilesh Murty; Goyal, Rahul; Schoelkopf, Bernhard; Fischer, Peer*Nature Communications, 2024, 15(1): 10692.
- Freely scalable and reconfigurable optical hardware for deep learningBernstein, Liane*; Sludds, Alexander*; Hamerly, Ryan; Sze, Vivienne; Emer, Joel; Englund, Dirk*Scientific Reports, 2021, 11(1): 3144.被引 43
- Spatially varying nanophotonic neural networksWei, Kaixuan; Li, Xiao; Froech, Johannes; Chakravarthula, Praneeth; Whitehead, James; Tseng, Ethan; Majumdar, Arka; Heide, Felix*Science Advances, 2024, 10(45): eadp0391.被引 1
