Bias-Reduced Uncertainty Estimation for Deep Neural Classifiers

2018/05/21 by Yonatan Geifman, Guy Uziel, Geifman, Yonatan +4 · 16 citations
Computer Science · Mathematics · #Adversarial Robustness in Machine Learning #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and Algorithms #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.1805.08206

Accepted to ICLR 2019

openalex publication_date 2018/05/21 · arxiv created 2019/04/24 · arxiv updated 2019/04/25 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We consider the problem of uncertainty estimation in the context of (non-Bayesian) deep neural classification. In this context, all known methods are based on extracting uncertainty signals from a trained network optimized to solve the classification problem at hand. We demonstrate that such techniques tend to introduce biased estimates for instances whose predictions are supposed to be highly confident. We argue that this deficiency is an artifact of the dynamics of training with SGD-like optimizers, and it has some properties similar to overfitting. Based on this observation, we develop an uncertainty estimation algorithm that selectively estimates the uncertainty of highly confident points, using earlier snapshots of the trained model, before their estimates are jittered (and way before they are ready for actual classification). We present extensive experiments indicating that the proposed algorithm provides uncertainty estimates that are consistently better than all known methods.

Cited by

• A Systematic Analysis of Out-of-Distribution Detection Under Representation and Training Paradigm Shifts
• SHRUG-FM: Reliability-Aware Foundation Models for Earth Observation
• What Does It Take to Build a Performant Selective Classifier?
• D-TPT: Dimensional Entropy Maximization for Calibrating Test-Time Prompt Tuning in Vision-Language Models
• Decoding Uncertainty: The Impact of Decoding Strategies for Uncertainty Estimation in Large Language Models
• Selective Classification Can Magnify Disparities Across Groups
• Spatially-Aware Evaluation of Segmentation Uncertainty
• Dense Uncertainty Estimation
• Revisiting Reweighted Risk for Calibration: AURC, Focal, and Inverse Focal Loss
• Aurora: Are Android Malware Classifiers Reliable and Stable under Distribution Shift?
• Suitability Filter: A Statistical Framework for Classifier Evaluation in Real-World Deployment Settings
• Asymmetric Duos: Sidekicks Improve Uncertainty
• Intra-class Patch Swap for Self-Distillation
• Confidence Calibration for Convolutional Neural Networks Using Structured Dropout
• OT Score: An OT based Confidence Score for Prototype-Assisted Source Free Unsupervised Domain Adaptation
• Reinforcement Learning Under Algorithmic Triage

Related

• Towards Principled Uncertainty Estimation for Deep Neural Networks
• Uncertainty Quantification in Deep Neural Networks through Statistical Inference on Latent Space
• Generalized Uncertainty of Deep Neural Networks: Taxonomy and Applications
• On Last-Layer Algorithms for Classification: Decoupling Representation from Uncertainty Estimation
• Dense Uncertainty Estimation