cross-entropy
A loss function commonly used in classification tasks, measuring the difference between the predicted probability distribution and the true distribution. It is an essential tool in training deep learning models to ensure better alignment of outputs with ground truth.
- A CLT for Polynomial GNNs on Community-Based Graphs
- Aligning Evaluation with Clinical Priorities: Calibration, Label Shift, and Error Costs
- Bigger, Regularized, Categorical: High-Capacity Value Functions are Efficient Multi-Task Learners
- Exploiting Vocabulary Frequency Imbalance in Language Model Pre-training
- Learning normalized image densities via dual score matching
- Rethinking Entropy in Test-Time Adaptation: The Missing Piece from Energy Duality
- Rethinking Fine-Tuning when Scaling Test-Time Compute: Limiting Confidence Improves Mathematical Reasoning