2023/01/13 by Rongzhen Zhao, Zhao, Rongzhen, Zhenzhi Wu +3
Computer Science · Mathematics · #Advanced Neural Network Applications #Algorithm #Artificial intelligence #Artificial neural network #Code (set theory) #Computation #Computer Vision and Pattern Recognition (cs.CV) #Computer science #Convolution (computer science) #Discrete mathematics #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Human Pose and Action Recognition #Kernel (algebra) #Mathematics #Parallel computing #Pattern recognition (psychology) #Theoretical computer science #Topology (electrical circuits)
paper · pdf · doi:10.48550/arxiv.2301.05440
published in arXiv (Cornell University) (Cornell University)
openalex publication_date 2023/01/13 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/08
Existing convolution techniques in artificial neural networks suffer from huge computation complexity, while the biological neural network works in a much more powerful yet efficient way. Inspired by the biological plasticity of dendritic topology and synaptic strength, our method, Learnable Heterogeneous Convolution, realizes joint learning of kernel shape and weights, which unifies existing handcrafted convolution techniques in a data-driven way. A model based on our method can converge with structural sparse weights and then be accelerated by devices of high parallelism. In the experiments, our method either reduces VGG16/19 and ResNet34/50 computation by nearly 5x on CIFAR10 and 2x on ImageNet without harming the performance, where the weights are compressed by 10x and 4x respectively; or improves the accuracy by up to 1.0% on CIFAR10 and 0.5% on ImageNet with slightly higher efficiency. The code will be available on www.github.com/Genera1Z/LearnableHeterogeneousConvolution.