Co-training for visual object recognition based on self-supervised models using a cross-entropy regularization

Mathematics

• Co-training 100%
• Object Recognition 73%
• Cross-entropy 69%
• Vision 49%
• Regularization 41%
• Model 15%
• Neural Networks 15%
• Learning 14%
• Output 12%
• Supervised Learning 11%
• Object 10%
• Simplicity 7%
• Filter 7%
• Experiment 6%
• Alternatives 4%
• Concepts 4%

Engineering & Materials Science

• Object recognition 50%
• Entropy 40%
• Labeled data 18%
• Deep learning 14%
• Supervised learning 8%
• Deep neural networks 7%
• Neural networks 4%
• Experiments 2%

Physics & Astronomy

• education 47%
• entropy 35%
• learning 26%
• output 9%
• filters 5%