PMID- 36607108
OWN - NLM
STAT- MEDLINE
DCOM- 20230110
LR  - 20230111
IS  - 1539-4522 (Electronic)
IS  - 1559-128X (Linking)
VI  - 61
IP  - 35
DP  - 2022 Dec 10
TI  - Optical computing powers graph neural networks.
PG  - 10471-10477
LID - 10.1364/AO.475991 [doi]
AB  - Graph-based neural networks have promising perspectives but are limited by 
      electronic bottlenecks. Our work explores the advantages of optical neural 
      networks in the graph domain. We propose an optical graph neural network (OGNN) 
      based on inverse-designed optical processing units (OPUs) to classify graphs with 
      optics. The OPUs, combined with two types of optical components, can perform 
      multiply-accumulate, matrix-vector multiplication, and matrix-matrix 
      multiplication operations. The proposed OGNN can classify typical non-Euclidean 
      MiniGCDataset graphs and successfully predict 1000 test graphs with 100% 
      accuracy. The OPU-formed optical-electrical graph attention network is also 
      scalable to handle more complex graph data, such as the Cora dataset, with 89.0% 
      accuracy.
FAU - Tang, Kaida
AU  - Tang K
FAU - Chen, Jianwei
AU  - Chen J
FAU - Jiang, Huaqing
AU  - Jiang H
FAU - Chen, Jun
AU  - Chen J
FAU - Jin, Shangzhong
AU  - Jin S
FAU - Hao, Ran
AU  - Hao R
LA  - eng
PT  - Journal Article
PL  - United States
TA  - Appl Opt
JT  - Applied optics
JID - 0247660
SB  - IM
MH  - *Algorithms
MH  - *Neural Networks, Computer
EDAT- 2023/01/07 06:00
MHDA- 2023/01/11 06:00
CRDT- 2023/01/06 09:06
PHST- 2023/01/06 09:06 [entrez]
PHST- 2023/01/07 06:00 [pubmed]
PHST- 2023/01/11 06:00 [medline]
AID - 522737 [pii]
AID - 10.1364/AO.475991 [doi]
PST - ppublish
SO  - Appl Opt. 2022 Dec 10;61(35):10471-10477. doi: 10.1364/AO.475991.