StructuredConvolutionsforEfficientNeuralNetwork Design
–Neural Information Processing Systems
In this work, we explore restricting the degrees of freedom of convolutional kernels by imposing a structure on them. This structure can be thought of as constructing the convolutional kernel by super-imposingseveral constant-heightkernels.
Neural Information Processing Systems
Feb-8-2026, 03:47:01 GMT
- Country:
- North America
- United States (0.04)
- Canada
- Quebec > Montreal (0.04)
- British Columbia > Metro Vancouver Regional District
- Vancouver (0.04)
- Europe
- Asia
- Middle East > Israel (0.04)
- China > Beijing
- Beijing (0.04)
- North America
- Technology: