Greedy Policy Search (GPS) is a simple algorithm that learns a policy for test-time data augmentation based on the predictive performance on a validation set. GPS starts with an empty policy and builds it in an iterative fashion. Each step selects a sub-policy that provides the largest improvement in calibrated log-likelihood of ensemble predictions and adds it to the current policy.
Source: Greedy Policy Search: A Simple Baseline for Learnable Test-Time AugmentationPaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Autonomous Driving | 21 | 5.16% |
Time Series Analysis | 15 | 3.69% |
Management | 13 | 3.19% |
Object Detection | 12 | 2.95% |
Retrieval | 12 | 2.95% |
Bayesian Optimization | 11 | 2.70% |
Decision Making | 10 | 2.46% |
Semantic Segmentation | 10 | 2.46% |
BIG-bench Machine Learning | 10 | 2.46% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |