Parameterized Temperature Scaling for Boosting the Expressive Power in Post-Hoc Uncertainty Calibration
About
We address the problem of uncertainty calibration and introduce a novel calibration method, Parametrized Temperature Scaling (PTS). Standard deep neural networks typically yield uncalibrated predictions, which can be transformed into calibrated confidence scores using post-hoc calibration methods. In this contribution, we demonstrate that the performance of accuracy-preserving state-of-the-art post-hoc calibrators is limited by their intrinsic expressive power. We generalize temperature scaling by computing prediction-specific temperatures, parameterized by a neural network. We show with extensive experiments that our novel accuracy-preserving approach consistently outperforms existing algorithms across a large number of model architectures, datasets and metrics.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Calibration | CIFAR-10 5000-sample half (test) | ECE0.0126 | 23 | |
| Post-hoc Calibration | CIFAR-10 5000-sample half seed 42 (test) | ECE0.0113 | 14 |