Kervolutional Neural Networks

By Chen Wang, Jianfei Yang, Lihua Xie, Junsong Yuan. 2019

71 citations recorded by the source corpus.

Convolutional neural networks (CNNs) have enabled the state-of-the-art performance in many computer vision tasks. However, little effort has been devoted to establishing convolution in non-linear space. Existing works mainly leverage on the activation layers, which can only provide point-wise non-linearity. To solve this problem, a new operation, kervolution (kernel convolution), is introduced to approximate complex behaviors of human perception systems leveraging on the kernel trick. It generalizes convolution, enhances the model capacity, and captures higher order interactions of features, via patch-wise kernel functions, but without introducing additional parameters. Extensive experiments show that kervolutional neural networks (KNN) achieve higher accuracy and faster convergence than baseline CNN.

Subfields: Computer Vision and Pattern Recognition, Industrial and Manufacturing Engineering

How researchers see this paper

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

In Computer Vision and Pattern Recognition

1. Sebastian A. Scherer
2. Yuheng Qiu
3. Yafei Hu
4. Wenshan Wang
5. Chen Wang
6. Lei Tai
7. Oncel Tuzel
8. Bo Li

In Industrial and Manufacturing Engineering

1. Yuheng Qiu
2. Sebastian A. Scherer
3. Chen Wang

Citations in the corpus

• Lifelong Graph Learning (2020, positive): “We can also construct other types of layers based on the topology G F , e.g ., extending convolution to kervolution [43] to improve the model expressivity or combining the pagerank algorithm [22] to further reduce feature over-smoothness.”
• Visual Memorability for Robotic Interestingness via Unsupervised Online Learning (2020, methodology): “capacity, reading accuracy is less aected by new knowledge, and results in safe and easier incremental learning. 6.3Translational Invariance Although CNN features are invariant to small translations [36], they still fail to recall memory when large translations occurs. To solve this problem, we intr”
• AdderNet: Do We Really Need Multiplications in Deep Learning? (2019, unknown): “[28] also discussed different metrics in deep networks.”