discriminability
The ability of a model to differentiate between different classes or outputs based on their features, directly affecting the model's performance on classification tasks.
- Continual Gaussian Mixture Distribution Modeling for Class Incremental Semantic Segmentation
- DGCBench: A Deep Graph Clustering Benchmark
- LayerNavigator: Finding Promising Intervention Layers for Efficient Activation Steering in Large Language Models
- Rethinking Entropy in Test-Time Adaptation: The Missing Piece from Energy Duality
- Synergy Between the Strong and the Weak: Spiking Neural Networks are Inherently Self-Distillers