mlsanityData to 7 Oct 2026

Parameterized Temperature Scaling for Boosting the Expressive Power in Post-Hoc Uncertainty Calibration

arXiv 2102.12182 ↗First posted 24 Feb 2021cs.LG cs.AIcited by 48

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. Cited by 48 papers in the map (data through 7 Oct 2026). Builds on 33 of its 269 references that are in the map. First posted on arXiv 24 Feb 2021.

Builds on

33 of 269 references.

References · 33

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  3. 2016Densely Connected Convolutional Networkscited by 6,691
  4. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
  5. 2017On Calibration of Modern Neural Networkscited by 4,860
  6. 2018MobileNetV2: Inverted Residuals and Linear Bottleneckscited by 4,773
  7. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,888
  8. 2016Xception: Deep Learning with Depthwise Separable Convolutionscited by 2,320
  9. 2015Variational Dropout and the Local Reparameterization Trickcited by 489
  10. 2019Measuring Calibration in Deep Learningcited by 443
  11. 2019Verified Uncertainty Calibrationcited by 307
  12. 2020Pitfalls of In-Domain Uncertainty Estimation and Ensembling in Deep Learningcited by 208
21 more
  1. 2020Mix-n-Match: Ensemble and Compositional Methods for Uncertainty Calibration in Deep Learningcited by 190
  2. 2019Evaluating model calibration in classificationcited by 165
  3. 2017Learning Proximal Operators: Using Denoising Networks for Regularizing Inverse Imaging Problemscited by 107
  4. 2020Calibration of Neural Networks using Splinescited by 97
  5. 2020Intra Order-preserving Functions for Calibration of Multi-Class Neural Networkscited by 65
  6. 2019Non-Parametric Calibration for Classificationcited by 57
  7. 2018Dirichlet-based Gaussian Processes for Large-scale Calibrated Classificationcited by 50
  8. 2018The Double Sphere Camera Modelcited by 50
  9. 2018What Makes Good Synthetic Training Data for Learning Disparity and Optical Flow Estimation?cited by 44
  10. 2017Associative Domain Adaptationcited by 32
  11. 2017Learning by Association - A versatile semi-supervised training method for neural networkscited by 19
  12. 2017Efficient Online Surface Correction for Real-time Large-Scale 3D Reconstructioncited by 12
  13. 2015Point-wise Map Recovery and Refinement from Functional Correspondencecited by 10
  14. 2016Spectral Decompositions using One-Homogeneous Functionalscited by 7
  15. 2017A Variational Approach to Shape-from-shading Under Natural Illuminationcited by 6
  16. 2015Collaborative Total Variation: A General Framework for Vectorial TV Modelscited by 5
  17. 2016A Combinatorial Solution to Non-Rigid 3D Shape-to-Image Matchingcited by 5
  18. 2017Dense Multi-view 3D-reconstruction Without Dense Correspondencescited by 5
  19. 2015Sublabel-Accurate Relaxation of Nonconvex Energiescited by 4
  20. 2020Improving Classifier Confidence using Lossy Label-Invariant Transformationscited by 4
  21. 2017Nonlinear Spectral Image Fusioncited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 9; 2024: 12; 2025: 11; 2026: 13.

48 papers.

Sort:

By area:

  1. 2025ConfTuner: Training Large Language Models to Express Their Confidence Verballycited by 37
  2. 2024Robust Calibration of Large Vision-Language Adapterscited by 28
  3. 2022Adaptive Temperature Scaling for Robust Calibration of Deep Neural Networkscited by 24
  4. 2023Proximity-Informed Calibration for Deep Neural Networkscited by 18
  5. 2023How to Fix a Broken Confidence Estimator: Evaluating Post-hoc Methods for Selective Classification with Deep Neural Networkscited by 15
  6. 2023ACLS: Adaptive and Conditional Label Smoothing for Network Calibrationcited by 15
  7. 2024Open-Vocabulary Calibration for Fine-tuned CLIPcited by 15
  8. 2023Beyond In-Domain Scenarios: Robust Density-Aware Calibrationcited by 14
  9. 2024Understanding and Mitigating Miscalibration in Prompt Tuning for Vision-Language Modelscited by 9
  10. 2024Feature Clipping for Uncertainty Calibrationcited by 8
  11. 2023Scaling of Class-wise Training Losses for Post-hoc Calibrationcited by 7
  12. 2024Optimizing Calibration by Gaining Aware of Prediction Correctnesscited by 4
36 more
  1. 2024Uncertainty Calibration with Energy Based Instance-wise Scaling in the Wild Datasetcited by 4
  2. 2024Accuracy-Preserving Calibration via Statistical Modeling on Probability Simplexcited by 3
  3. 2024Decoupling of neural network calibration measurescited by 3
  4. 2024Consistency Calibration: Improving Uncertainty Calibration via Consistency among Perturbed Neighborscited by 3
  5. 2024Ordinal-Aware Calibration for Ordinal Classificationcited by 3
  6. 2025Enhancing Out-of-Distribution Detection with Extended Logit Normalizationcited by 3
  7. 2025Sample Margin-Aware Recalibration of Temperature Scalingcited by 3
  8. 2023Sensitivity analysis of AI-based algorithms for autonomous driving on optical wavefront aberrations induced by the windshieldcited by 2
  9. 2025h-calibration: Rethinking Classifier Recalibration with Probabilistic Error-Bounded Objectivecited by 2
  10. 2026Quantile Adaptive Temperature Scaling for Confidence Calibrationcited by 2
  11. 2023Perception and Semantic Aware Regularization for Sequential Confidence Calibrationcited by 1
  12. 2023IW-GAE: Importance Weighted Group Accuracy Estimation for Improved Calibration and Model Selection in Unsupervised Domain Adaptationcited by 1
  13. 2024Tilt and Average : Geometric Adjustment of the Last Layer for Recalibrationcited by 1
  14. 2025Classifier Ensemble for Efficient Uncertainty Calibration of Deep Neural Networks for Image Classificationcited by 1
  15. 2025Adaptive Cumulative Mass Calibration with Conformal Predictioncited by 1
  16. 2025Beyond Overconfidence: Foundation Models Redefine Calibration in Deep Neural Networkscited by 1
  17. 2025Probability Bounding: Post-Hoc Calibration via Box-Constrained Softmaxcited by 1
  18. 2025Instance-Wise Monotonic Calibration by Constrained Transformationcited by 1
  19. 2025Uncertainty Calibration of Multi-Label Bird Sound Classifierscited by 1
  20. 2026The Well-Tempered Classifier: Some Elementary Properties of Temperature Scalingcited by 1
  21. 2026Probing Routing-Conditional Calibration in Attention-Residual Transformerscited by 1
  22. 2022CHALLENGER: Training with Attribution Mapscited by 0
  23. 2022Estimating Classification Confidence Using Kernel Densitiescited by 0
  24. 2023Uncertainty Estimation based on Geometric Separationcited by 0
  25. 2024Optical aberrations in autonomous driving: Physics-informed parameterized temperature scaling for neural network uncertainty calibrationcited by 0
  26. 2025DATS: Distance-Aware Temperature Scaling for Calibrated Class-Incremental Learningcited by 0
  27. 2026Enhancing the Reliability of Medical AI through Expert-guided Uncertainty Modelingcited by 0
  28. 2026Adaptive Calibration for Fair and Performant Facial Recognitioncited by 0
  29. 2026Recovering Stranded Discrimination in Knowledge Tracing: Per-Item Bias Correction via Empirical-Bayes Shrinkagecited by 0
  30. 2026Eigenvalue Calibration for Semantic Embeddings of Large Language Modelscited by 0
  31. 2026Rethinking Detection Calibration: A Coordinate and Direction Perspectivecited by 0
  32. 2026CRS-Triage: Confidence- and Reliability-Aware Selective Triage under Incomplete Clinical Evidencecited by 0
  33. 2026Invertible Logits Transformation for Accuracy-Preserving Post-Hoc Uncertainty Calibrationcited by 0
  34. 2026CARD: Calibration via Agreement in Reverse Diffusion for Out-of-Domain MRI Segmentationcited by 0
  35. 2026Let Confidence Change, Not the Prediction: Prediction-Preserving Repair for Post-hoc Calibrationcited by 0
  36. 2026Auditing Routing Entropy as an Uncertainty Signal in Attention-Residual Transformerscited by 0