The 69th JSAP Spring Meeting 2022

Presentation information

Oral presentation

3 Optics and Photonics » 3.15 Silicon photonics and integrated photonics

[22p-E303-1~13] 3.15 Silicon photonics and integrated photonics

Tue. Mar 22, 2022 1:30 PM - 5:15 PM E303 (E303)

Yuya Shoji(Tokyo Tech), Tadashi Murao(Mitsubishi Electric), Kondo Keisuke(Utsunomiya Univ.)

3:00 PM - 3:15 PM

[22p-E303-6] Study on Robust Integrated Optical Neural Network using Multi-Plane Light Conversion

〇(DC)Ryota Tanomura1, Keigo Mizukami1, Rui Tang1, Go Soma1, Takuo Tanemura1, Yoshiaki Nakano1 (1.The Univ. Tokyo)

Keywords:optical neural network, optical unitary converter

Optical neural networks have been attracting attention as low-power and low-latency hardware. In particular, optical neural networks using integrated optical unitary transducers have attracted much attention. Still, conventional integrated optical unitary transducers are sensitive to errors, making it difficult to scale them up. In this study, we show that it is possible to realize robust hardware by using optical unitary converters based on a different method called multi-plane light conversion.