FixMatch is an algorithm that first generates pseudo-labels using the model's predictions on weakly-augmented unlabeled images. For a given image, the pseudo-label is only retained if the model produces a high-confidence prediction. The model is then trained to predict the pseudo-label when fed a strongly-augmented version of the same image.
Description from: FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
Image credit: FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence
Source: FixMatch: Simplifying Semi-Supervised Learning with Consistency and ConfidencePaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Pseudo Label | 18 | 15.38% |
Semi-Supervised Image Classification | 16 | 13.68% |
Image Classification | 8 | 6.84% |
Semantic Segmentation | 7 | 5.98% |
Semi-Supervised Semantic Segmentation | 5 | 4.27% |
Active Learning | 3 | 2.56% |
Domain Adaptation | 3 | 2.56% |
Classification | 3 | 2.56% |
Unsupervised Domain Adaptation | 2 | 1.71% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |