Greedy Modality Selection via Approximate Submodular Maximization

Keyphrases

• Greedy 100%
• Utility Measures 100%
• Submodular Maximization 100%
• Modality Selection 100%
• Multimodal Learning 66%
• Optimization Problem 33%
• Heterogeneous Data Sources 33%
• Multimodal Data 33%
• Value-based 33%
• Memory Constraints 33%
• Computational Constraints 33%
• Further Training 33%
• Shapley Value 33%
• Redundant Information 33%
• MNIST 33%
• Feature Importance Score 33%
• Approximate Submodularity 33%
• Synthetic Patch 33%
• Complementary Modalities 33%
• Study Modality 33%

Computer Science

• Multimodal Learning 100%
• Theoretical Framework 50%
• Efficient Algorithm 50%
• Optimization Problem 50%
• Multimodality 50%
• Switching Frequency 50%