A regularization criterion that, differently from dropout and its variants, is deterministic rather than random. It grounds on the empirical evidence that feature descriptors with larger L2-norm and highly-active nodes are strongly correlated to confident class predictions. Thus, the criterion guides towards dropping a percentage of the most active nodes of the descriptors, proportionally to the estimated class probability
Paper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Decision Making | 4 | 8.51% |
Visual Reasoning | 3 | 6.38% |
Semantic Segmentation | 3 | 6.38% |
Time Series Analysis | 2 | 4.26% |
Continual Learning | 1 | 2.13% |
Incremental Learning | 1 | 2.13% |
Benchmarking | 1 | 2.13% |
Language Modelling | 1 | 2.13% |
Large Language Model | 1 | 2.13% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |