Neural Architecture Search
Neural Architecture Search (NAS) is a technique for automating the design of Artificial Neural Networks (ANN). Its goal is to automatically explore and optimize network structures through algorithms to discover more efficient and complex model architectures, thereby enhancing the performance and effectiveness of machine learning tasks. The application value of NAS lies in reducing the time cost associated with manual parameter tuning, improving the efficiency and accuracy of model design.
ImageNet
NAT-M4
NAS-Bench-201, ImageNet-16-120
Shapley-NAS
CIFAR-10
NAT-M4
NAS-Bench-201, CIFAR-100
Shapley-NAS
NAS-Bench-201, CIFAR-10
GenNAS
CIFAR-10 Image Classification
EEEA-Net-C (b=5)+ CO
CIFAR-100
DNA-c
NATS-Bench Topology, ImageNet16-120
GreenMachine-1
NATS-Bench Topology, CIFAR-10
NATS-Bench Topology, CIFAR-100
Food-101
Balanced Mixture
NAS-Bench-101
FireFly
NATS-Bench Size, CIFAR-10
BossNAS
NATS-Bench Size, CIFAR-100
CINIC-10
NAT-M4
DTD
NAT-M4
FGVC Aircraft
NAT-M4
NATS-Bench Size, ImageNet16-120
Oxford 102 Flowers
NAT-M4
Oxford-IIIT Pet Dataset
NAT-M4
Stanford Cars
NAT-M4
STL-10
NAT-M4
NAS-Bench-201
Improved FireFly Algorithme
LIDC-IDRI
NASLung (ours)
MNIST
NAS-Bench-301
DiNAS