Instance Wise Uncertainty

Instance-wise uncertainty quantification aims to estimate the reliability of individual predictions made by machine learning models, addressing the limitations of traditional methods that provide only aggregate uncertainty measures. Current research focuses on developing methods to accurately estimate this uncertainty across diverse tasks, including image segmentation, regression, and extreme multi-label classification, employing techniques like ensemble methods, Gaussian processes, and adaptations of gradient boosting. This research is crucial for improving the trustworthiness and reliability of AI systems in high-stakes applications like medical image analysis and protein engineering, where understanding the confidence of individual predictions is paramount for safe and effective deployment.

Papers

October 16, 2024

TPFL: A Trustworthy Personalized Federated Learning Framework via Subjective Logic
Jinqian Chen, Jihua Zhu
Native Robustness Personalized Federated Learning Instance Wise Uncertainty Subjective Logic

July 17, 2024

Instance-wise Uncertainty for Class Imbalance in Semantic Segmentation
Luís Almeida, Inês Dutra, Francesco Renna
Semantic Segmentation Uncertainty Estimation Class Imbalance Pixel Correlation Instance Wise Uncertainty

July 8, 2024

Interpretability of Uncertainty: Exploring Cortical Lesion Segmentation in Multiple Sclerosis
Nataliia Molchanova, Alessandro Cagol, Pedro M. Gordaliza, Mario Ocampo-Pineda, Po-Jui Lu, Matthias Weigel, Xinjie Chen, Adrien Depeursinge, Cristina Granziera, Henning Müller, Meritxell Bach Cuadra
Medical Image Segmentation Inherent Interpretability Uncertainty Quantification High Uncertainty Anticipation Lesion Segmentation Multiple Sclerosis Brain Lesion Instance Wise Uncertainty

May 7, 2024

Weakly-Supervised Residual Evidential Learning for Multi-Instance Uncertainty Estimation
Pei Liu, Luping Ji
Uncertainty Estimation Multiple Instance Learning Predictive Uncertainty Instance Learning Evidential Learning Instance Wise Uncertainty

April 9, 2024

Kermut: Composite kernel regression for protein variant effects
Peter Mørch Groth, Mads Herbert Kerrn, Lars Olsen, Jesper Salomon, Wouter Boomsma
Kernel Regression Non Structural Protein Protein Engineering Instance Wise Uncertainty Sequence Similarity Variant Effect Prediction

February 21, 2024

Private Gradient Descent for Linear Regression: Tighter Error Bounds and Instance-Specific Uncertainty Estimation
Gavin Brown, Krishnamurthy Dvijotham, Georgina Evans, Daogao Liu, Adam Smith, Abhradeep Thakurta
Sample Complexity Linear Regression Private Algorithm Error Bound Private Gradient Square Loss Instance Wise Uncertainty

September 27, 2023

Uncertainty Quantification via Neural Posterior Principal Components
Elias Nehme, Omer Yair, Tomer Michaeli
Uncertainty Quantification Image Restoration Model Uncertainty Visualization Instance Wise Uncertainty Pixel Variance

May 5, 2023

Uncertainty-Aware Bootstrap Learning for Joint Extraction on Distantly-Supervised Data
Yufei Li, Xiao Yu, Yanchi Liu, Haifeng Chen, Cong Liu
Entity Pair Distant Supervision Large Scale Semisupervised Bootstrapping Joint Extraction Instance Wise Uncertainty

October 18, 2022

Uncertainty in Extreme Multi-label Classification
Jyun-Yu Jiang, Wei-Cheng Chang, Jiong Zhong, Cho-Jui Hsieh, Hsiang-Fu Yu
Uncertainty Quantification High Uncertainty Anticipation Supervised Baseline Extreme Multi Label Instance Wise Uncertainty

May 23, 2022

Instance-Based Uncertainty Estimation for Gradient-Boosted Regression Trees
Jonathan Brophy, Daniel Lowd
Gradient Boosted Tree Instance Wise Uncertainty Ensemble Kernel

Instance Wise Uncertainty

Papers

TPFL: A Trustworthy Personalized Federated Learning Framework via Subjective Logic

Instance-wise Uncertainty for Class Imbalance in Semantic Segmentation

Interpretability of Uncertainty: Exploring Cortical Lesion Segmentation in Multiple Sclerosis

Weakly-Supervised Residual Evidential Learning for Multi-Instance Uncertainty Estimation

Kermut: Composite kernel regression for protein variant effects

Private Gradient Descent for Linear Regression: Tighter Error Bounds and Instance-Specific Uncertainty Estimation

Uncertainty Quantification via Neural Posterior Principal Components

Uncertainty-Aware Bootstrap Learning for Joint Extraction on Distantly-Supervised Data

Uncertainty in Extreme Multi-label Classification

Instance-Based Uncertainty Estimation for Gradient-Boosted Regression Trees