Evidential Deep Learning to Quantify Classification Uncertainty

Abstract

Deterministic neural nets have been shown to learn effective predictors on a wide range of machine learning problems. However, as the standard approach is to train the network to minimize a prediction loss, the resultant model remains ignorant to its prediction confidence. Orthogonally to Bayesian neural nets that indirectly infer prediction uncertainty through weight uncertainties, we propose explicit modeling of the same using the theory of subjective logic. By placing a Dirichlet distribution on the class probabilities, we treat predictions of a neural net as subjective opinions and learn the function that collects the evidence leading to these opinions by a deterministic neural net from data. The resultant predictor for a multi-class classification problem is another Dirichlet distribution whose parameters are set by the continuous output of a neural net. We provide a preliminary analysis on how the peculiarities of our new loss function drive improved uncertainty estimation. We observe that our method achieves unprecedented success on detection of out-of-distribution queries and endurance against adversarial perturbations.

Cited by in corpus (99)

• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• Aleatoric and Epistemic Uncertainty in Machine Learning: An Introduction to Concepts and Methods
• Uncertainty Quantification in Scientific Machine Learning: Methods, Metrics, and Comparisons
• A Review and Comparative Study on Probabilistic Object Detection in Autonomous Driving
• Uncertainty Quantification in Machine Learning for Engineering Design and Health Prognostics: A Tutorial
• Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance Awareness
• Stochastic Segmentation Networks: Modelling Spatially Correlated Aleatoric Uncertainty
• AI Security for Geoscience and Remote Sensing: Challenges and Future Trends
• Reverse KL-Divergence Training of Prior Networks: Improved Uncertainty and Adversarial Robustness
• Posterior Network: Uncertainty Estimation without OOD Samples via Density-Based Pseudo-Counts
• Medical Imaging and Machine Learning
• Uncertainty Aware Semi-Supervised Learning on Graph Data
• Trusted Multi-View Classification
• Uncertainty-Induced Transferability Representation for Source-Free Unsupervised Domain Adaptation
• Reliability and Interpretability in Science and Deep Learning
• Uncertainty Aware Training to Improve Deep Learning Model Calibration for Classification of Cardiac MR Images
• Being Bayesian, Even Just a Bit, Fixes Overconfidence in ReLU Networks
• The Unreasonable Effectiveness of Deep Evidential Regression
• Uncertainty Sets for Image Classifiers using Conformal Prediction
• Curriculum learning for improved femur fracture classification: scheduling data with prior knowledge and uncertainty
• Information Aware Max-Norm Dirichlet Networks for Predictive Uncertainty Estimation
• Increasing the Confidence of Deep Neural Networks by Coverage Analysis
• Uncertainty-Aware Reliable Text Classification
• Operational Calibration: Debugging Confidence Errors for DNNs in the Field
• CertainNet: Sampling-free Uncertainty Estimation for Object Detection
• Reliable Neural Networks for Regression Uncertainty Estimation
• Prior and Posterior Networks: A Survey on Evidential Deep Learning Methods For Uncertainty Estimation
• Regression Prior Networks
• Uncertainty-Aware Machine Translation Evaluation
• Dirichlet-Based Prediction Calibration for Learning with Noisy Labels
• UMat: Uncertainty-Aware Single Image High Resolution Material Capture
• A Variational Dirichlet Framework for Out-of-Distribution Detection
• Improved Adversarial Robustness by Reducing Open Space Risk via Tent Activations
• Semi-supervised Active Learning for Instance Segmentation via Scoring Predictions
• RSEA-MVGNN: Multi-View Graph Neural Network with Reliable Structural Enhancement and Aggregation
• Deep Evidential Regression
• Quantifying Deep Learning Model Uncertainty in Conformal Prediction
• Estimating the Uncertainty in Emotion Attributes using Deep Evidential Regression
• A Unified Benchmark for the Unknown Detection Capability of Deep Neural Networks
• Being Bayesian about Categorical Probability
• Deep Evidential Learning for Radiotherapy Dose Prediction
• Evidential Deep Learning for Open Set Action Recognition
• Fast Predictive Uncertainty for Classification with Bayesian Deep Networks
• Graph Posterior Network: Bayesian Predictive Uncertainty for Node Classification
• Accurate Passive Radar via an Uncertainty-Aware Fusion of Wi-Fi Sensing Data
• Towards neural networks that provably know when they don't know
• Computing a human-like reaction time metric from stable recurrent vision models
• Vision-Based Autonomous Car Racing Using Deep Imitative Reinforcement Learning
• Explainable AI for Intelligence Augmentation in Multi-Domain Operations
• Using Subjective Logic to Estimate Uncertainty in Multi-Armed Bandit Problems
• Robust Semantic Segmentation with Superpixel-Mix
• Uncertainty Quantification for Deep Context-Aware Mobile Activity Recognition and Unknown Context Discovery
• Mean-Field Approximation to Gaussian-Softmax Integral with Application to Uncertainty Estimation
• Evaluating Robustness of Predictive Uncertainty Estimation: Are Dirichlet-based Models Reliable?
• Provably Robust Detection of Out-of-distribution Data (almost) for free
• No Surprises: Training Robust Lung Nodule Detection for Low-Dose CT Scans by Augmenting with Adversarial Attacks
• Multi-Loss Sub-Ensembles for Accurate Classification with Uncertainty Estimation
• Deep Dirichlet uncertainty for unsupervised out-of-distribution detection of eye fundus photographs in glaucoma screening
• Graceful Degradation and Related Fields
• NSL: Hybrid Interpretable Learning From Noisy Raw Data
• DREAMS: A python framework for Training Deep Learning Models on EEG Data with Model Card Reporting for Medical Applications
• Fine-grained Uncertainty Modeling in Neural Networks
• MambaEviScrib: Mamba and Evidence-Guided Consistency Enhance CNN Robustness for Scribble-Based Weakly Supervised Ultrasound Image Segmentation
• Learning Partially Known Stochastic Dynamics with Empirical PAC Bayes
• Evidential Sparsification of Multimodal Latent Spaces in Conditional Variational Autoencoders
• Uncertainty estimations methods for a deep learning model to aid in clinical decision-making -- a clinician's perspective
• Functional Space Variational Inference for Uncertainty Estimation in Computer Aided Diagnosis
• Towards Interactive Reinforcement Learning with Intrinsic Feedback
• Trustworthy Long-Tailed Classification
• Visual hallucination detection in large vision-language models via evidential conflict
• Being a Bit Frequentist Improves Bayesian Neural Networks
• Classification and Uncertainty Quantification of Corrupted Data using Semi-Supervised Autoencoders
• Identifying Incorrect Classifications with Balanced Uncertainty
• Prior Activation Distribution (PAD): A Versatile Representation to Utilize DNN Hidden Units
• Efficient and Robust LiDAR-Based End-to-End Navigation
• Uncertainty Propagation in Node Classification
• Online Black-Box Confidence Estimation of Deep Neural Networks
• Improving Explainable Object-induced Model through Uncertainty for Automated Vehicles
• Uncertainty quantification for White Matter Hyperintensity segmentation detects silent failures and improves automated Fazekas quantification
• Query-based Hard-Image Retrieval for Object Detection at Test Time
• Learning a metacognition for object perception
• Deep Out-of-Distribution Uncertainty Quantification via Weight Entropy Maximization
• Robust Partial-Label Learning by Leveraging Class Activation Values
• APPLI: Adaptive Planner Parameter Learning From Interventions
• Towards human-agent knowledge fusion (HAKF) in support of distributed coalition teams
• Deep Classifiers with Label Noise Modeling and Distance Awareness
• Maximizing BCI Human Feedback using Active Learning
• FADEL: Uncertainty-aware Fake Audio Detection with Evidential Deep Learning
• FROB: Few-shot ROBust Model for Classification and Out-of-Distribution Detection
• Expectation-Maximization Regularized Deep Learning for Weakly Supervised Tumor Segmentation for Glioblastoma
• Adversarial Ranking Attack and Defense
• Uncertainty Estimation in Deep Neural Networks for Point Cloud Segmentation in Factory Planning
• Multivariate Deep Evidential Regression
• Diagnostic Uncertainty Calibration: Towards Reliable Machine Predictions in Medical Domain
• Evidential Turing Processes
• Robust Vision-Based Cheat Detection in Competitive Gaming
• Classifying different types of solar wind plasma with uncertainty estimations using machine learning
• Boosting Cross-Lingual Transfer via Self-Learning with Uncertainty Estimation
• Generalized Zero-Shot Learning for Point Cloud Segmentation with Evidence-Based Dynamic Calibration