Fine-Tuning DARTS for Image Classification
This work addresses performance issues in neural architecture search for image classification, offering incremental improvements in accuracy with a trade-off in parameters.
The paper tackles the problem of inferior performance in Differential Architecture Search (DARTS) due to approximations by proposing fine-tuning with fixed operations, resulting in improved top-1 accuracy on datasets like Fashion-MNIST, CompCars, and MIO-TCD by 0.56%, 0.50%, and 0.39% respectively compared to state-of-the-art approaches.
Neural Architecture Search (NAS) has gained attraction due to superior classification performance. Differential Architecture Search (DARTS) is a computationally light method. To limit computational resources DARTS makes numerous approximations. These approximations result in inferior performance. We propose to fine-tune DARTS using fixed operations as they are independent of these approximations. Our method offers a good trade-off between the number of parameters and classification accuracy. Our approach improves the top-1 accuracy on Fashion-MNIST, CompCars, and MIO-TCD datasets by 0.56%, 0.50%, and 0.39%, respectively compared to the state-of-the-art approaches. Our approach performs better than DARTS, improving the accuracy by 0.28%, 1.64%, 0.34%, 4.5%, and 3.27% compared to DARTS, on CIFAR-10, CIFAR-100, Fashion-MNIST, CompCars, and MIO-TCD datasets, respectively.