Weight-sharing neural architecture search: A battle to shrink the optimization gap
L ** out of performance collapse without indicators
Despite the fast development of differentiable architecture search (DARTS), it suffers from
long-standing performance instability, which extremely limits its application. Existing …
long-standing performance instability, which extremely limits its application. Existing …
Progressive darts: Bridging the optimization gap for nas in the wild
With the rapid development of neural architecture search (NAS), researchers found powerful
network architectures for a wide range of vision tasks. Like the manually designed …
network architectures for a wide range of vision tasks. Like the manually designed …
Direct differentiable augmentation search
Data augmentation has been an indispensable tool to improve the performance of deep
neural networks, however the augmentation can hardly transfer among different tasks and …
neural networks, however the augmentation can hardly transfer among different tasks and …
Fine-tuning darts for image classification
Neural Architecture Search (NAS) has gained attraction due to superior classification
performance. Differential Architecture Search (DARTS) is a computationally light method. To …
performance. Differential Architecture Search (DARTS) is a computationally light method. To …
Zarts: On zero-order optimization for neural architecture search
Differentiable architecture search (DARTS) has been a popular one-shot paradigm for NAS
due to its high efficiency. It introduces trainable architecture parameters to represent the …
due to its high efficiency. It introduces trainable architecture parameters to represent the …
ALADA: A lite automatic data augmentation framework for industrial defect detection
Industrial defect detection is a critical and challenging task in the quality control of
manufacturing production. Competent in feature extraction and pattern recognition, deep …
manufacturing production. Competent in feature extraction and pattern recognition, deep …
Efficient evaluation methods for neural architecture search: A survey
Neural Architecture Search (NAS) has received increasing attention because of its
exceptional merits in automating the design of Deep Neural Network (DNN) architectures …
exceptional merits in automating the design of Deep Neural Network (DNN) architectures …