Active privileged learning of human activities from weakly labeled samples

Keyphrases

• Training Data 100%
• Labeled Sample 100%
• Weak Labels 100%
• Human Activities 100%
• Privileged Learning 100%
• Low Quality Data 50%
• Testing Data 50%
• Unlabeled Data 50%
• Labeled Data 50%
• Probabilistic Approach 50%
• Active Learning 50%
• Active Learning Approach 50%
• Human Effort 50%
• Pool-based 50%
• Publicly Available Dataset 50%
• Decision Boundary 50%
• Learning Using Privileged Information 50%
• Data Collection Process 50%
• Complex Human Activity 50%
• Data Annotation 50%
• Privileged Information Learning 50%
• Human Activity Recognition Systems 50%
• Query Inference 50%

Computer Science

• Active Learning 100%
• Experimental Result 50%
• Learning Approach 50%
• Unlabeled Data 50%
• Annotation 50%
• Training Data 50%
• Probabilistic Approach 50%
• Decision Boundary 50%
• Recognition System 50%
• Data Collection Process 50%
• Human Activity Recognition 50%