UPANets: Learning from the Universal Pixel Attention Neworks

Mathematics

• Pixel 100%
• Learning 75%
• Resources 58%
• Stacking 56%
• Computer Vision 49%
• Neural Networks 38%
• Receptive Field 31%
• Transformer 27%
• Backbone 26%
• Hardware 25%
• Expand 21%
• Customers 21%
• Character 15%
• Experiment 15%
• Series 12%
• Performance 12%
• Model 6%

Engineering & Materials Science

• Graphics processing unit 77%
• Convolutional neural networks 75%
• Computer vision 69%
• Pixels 62%
• Deep neural networks 40%
• Computer hardware 24%
• Experiments 14%

Physics & Astronomy

• learning 90%
• pixels 67%
• resources 48%
• computer vision 42%
• ramps 17%
• transformers 17%
• hardware 15%
• performance 7%