Fast Neural Network Adaptation via Parameter Remapping and Architecture Search

By Jiemin Fang, Yuzhu Sun, Kangjian Peng, Qian Zhang, Yuan Li, Wenyu Liu, Xinggang Wang. 2020

36 citations recorded by the source corpus.

Deep neural networks achieve remarkable performance in many computer vision tasks. Most state-of-the-art~(SOTA) semantic segmentation and object detection approaches reuse neural network architectures designed for image classification as the backbone, commonly pre-trained on ImageNet. However, performance gains can be achieved by designing network architectures specifically for detection and segmentation, as shown by recent neural architecture search (NAS) research for detection and segmentation. One major challenge though, is that ImageNet pre-training of the search space representation (a.k.a. super network) or the searched networks incurs huge computational cost. In this paper, we propose a Fast Neural Network Adaptation (FNA) method, which can adapt both the architecture and parameters of a seed network (e.g. a high performing manually designed backbone) to become a network with different depth, width, or kernels via a Parameter Remapping technique, making it possible to utilize NAS for detection/segmentation tasks a lot more efficiently. In our experiments, we conduct FNA on MobileNetV2 to obtain new networks for both segmentation and detection that clearly out-perform existing networks designed both manually and by NAS. The total computation cost of FNA is significantly less than SOTA segmentation/detection NAS approaches: 1737$\times$ less than DPC, 6.8$\times$ less than Auto-DeepLab and 7.4$\times$ less than DetNAS. The code is available at https://github.com/JaminFong/FNA .

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Efficiency & Compression, Neural Architecture Search & AutoML

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Xinggang Wang
2. Wenyu Liu
3. Qian Zhang
4. Po-Yao (Bernie) Huang
5. Xiaojun Chang
6. Zhihui Li
7. Yuan Li
8. Yuzhu Sun

In Computer Vision and Pattern Recognition

1. Wenyu Liu
2. Xinggang Wang
3. Qian Zhang
4. Xiaojun Chang
5. Jiemin Fang
6. Yuan Li
7. Po-Yao (Bernie) Huang
8. Zhihui Li

In Efficiency & Compression

1. Xiaojun Chang
2. Qian Zhang
3. Wenyu Liu
4. Yuan Li
5. Yuzhu Sun
6. Xinggang Wang
7. Zhihui Li
8. Jiemin Fang

In Neural Architecture Search & AutoML

1. Xiaojun Chang
2. Xin Wang
3. Po-Yao (Bernie) Huang
4. Xiaojiang Chen
5. Yun Xiao
6. Zhihui Li
7. Pengzhen Ren

Citations in the corpus

• A Comprehensive Survey of Neural Architecture Search (2020, methodology): “Fast Neural Network Adaptation (FNA) [59] proposes a method that can adapt the architecture and parameters of a network to new tasks at almost zero cost.”
• Densely Connected Search Space for More Flexible Neural Architecture Search (2019, methodology): “FNA [15] proposes to adapt the neural network to new tasks with low cost by a parameter remap-ping mechanism and differentiable NAS.”