Max-plus operators applied to filter selection and model pruning in neural networks
From MaRDI portal
Publication:2189383
Abstract: Following recent advances in morphological neural networks, we propose to study in more depth how Max-plus operators can be exploited to define morphological units and how they behave when incorporated in layers of conventional neural networks. Besides showing that they can be easily implemented with modern machine learning frameworks , we confirm and extend the observation that a Max-plus layer can be used to select important filters and reduce redundancy in its previous layer, without incurring performance loss. Experimental results demonstrate that the filter selection strategy enabled by a Max-plus is highly efficient and robust, through which we successfully performed model pruning on different neural network architectures. We also point out that there is a close connection between Maxout networks and our pruned Max-plus networks by comparing their respective characteristics. The code for reproducing our experiments is available online.
Recommendations
- Sparse approximate solutions to max-plus equations
- A simple neural network pruning algorithm with application to filter synthesis
- Fast convex pruning of deep neural networks
- Morphological Perceptrons: Geometry and Training Algorithms
- MorphoActivation: generalizing ReLU activation function by mathematical morphology
Cited in
(11)- Approximating morphological operators with part-based representations learned by asymmetric auto-encoders
- Some open questions on morphological operators and representations in the deep learning era. A personal vision
- On some associations between mathematical morphology and artificial intelligence
- Going beyond p-convolutions to learn grayscale morphological operators
- Sparse approximate solutions to max-plus equations
- Toward a sparsity theory on weighted lattices
- Learning grayscale mathematical morphology with smooth morphological layers
- Learnable Empirical Mode Decomposition based on Mathematical Morphology
- MorphoActivation: generalizing ReLU activation function by mathematical morphology
- Logarithmic morphological neural nets robust to lighting variations
- Morphological neural networks: a review
This page was built for publication: Max-plus operators applied to filter selection and model pruning in neural networks
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2189383)
