Centroid Estimation with Guaranteed Efficiency: A General Framework for Weakly Supervised Learning

Keyphrases

• Unsupervised Learning 100%
• Centroid Measurement 100%
• Centroid 40%
• Abusive Supervision 40%
• Statistical Efficiency 40%
• Minimum Variance 20%
• Learning Problems 20%
• Loss Function 20%
• Semi-supervised Learning 20%
• Labeled Data 20%
• Benchmark Dataset 20%
• Unbiased Estimate 20%
• Multiple Instance Learning 20%
• Hinge Loss 20%
• Auxiliary Dataset 20%
• Centroid-based 20%
• Squared Loss 20%
• Positive-unlabeled Learning 20%
• Noisy Label Learning 20%
• Risk Estimator 20%
• Inexact Supervision 20%

Computer Science

• Supervised Learning 100%
• Statistical Efficiency 66%
• Learning Problem 33%
• Multiple Instance 33%
• Instance Learning 33%
• Semisupervised Learning 33%
• Risk Estimator 33%
• Experimental Result 33%