Learning Sample Difficulty from Pre-trained Models for Reliable Prediction
–Neural Information Processing Systems
Equipped with the knowledge of sample difficulty, we further propose to use it for regularizing the prediction confidence.
Neural Information Processing Systems
Oct-8-2025, 16:29:51 GMT
- Country:
- North America > United States
- Europe
- Slovenia > Drava
- Municipality of Benedikt > Benedikt (0.04)
- Germany > Baden-Württemberg
- Tübingen Region > Tübingen (0.04)
- Slovenia > Drava
- Asia > China
- Shanghai > Shanghai (0.04)
- Shandong Province (0.04)
- Beijing > Beijing (0.04)
- Technology:
- Information Technology
- Sensing and Signal Processing > Image Processing (0.93)
- Data Science > Data Mining (0.93)
- Artificial Intelligence
- Vision (1.00)
- Natural Language (0.93)
- Machine Learning
- Performance Analysis > Accuracy (1.00)
- Neural Networks (0.70)
- Information Technology