Class Uncertainty

Class uncertainty, in the context of machine learning, refers to the quantification and mitigation of a model's confidence in its predictions, particularly concerning the ambiguity inherent in class boundaries or the presence of imbalanced datasets. Current research focuses on developing novel metrics beyond simple class cardinality to better capture class uncertainty, incorporating these metrics into existing class imbalance mitigation techniques, and leveraging uncertainty estimates for improved model training and generalization, often employing evidential deep learning and techniques like Monte Carlo Dropout. Understanding and addressing class uncertainty is crucial for building more reliable and robust machine learning systems, particularly in high-stakes applications where accurate prediction confidence is paramount.

Papers

August 21, 2024

Diagnosing and Remedying Knowledge Deficiencies in LLMs via Label-free Curricular Meaningful Learning
Kai Xiong, Xiao Ding, Li Du, Jiahao Ying, Ting Liu, Bing Qin, Yixin Cao
Large Language Model Curriculum Learning Knowledge Gap Class Uncertainty

November 23, 2023

Class Uncertainty: A Measure to Mitigate Class Imbalance
Z. S. Baltaci, K. Oksuz, S. Kuzucu, K. Tezoren, B. K. Konar, A. Ozkan, E. Akbas, S. Kalkan
Class Imbalance Cardinality Estimation Class Uncertainty

November 19, 2023

Evidential Uncertainty Quantification: A Variance-Based Perspective
Ruxiao Duan, Brian Caffo, Harrison X. Bai, Haris I. Sair, Craig Jones
Evidential Deep Learning Classification Uncertainty Evidential Uncertainty Variance Aware Class Uncertainty

October 30, 2023

Uncertainty-guided Boundary Learning for Imbalanced Social Event Detection
Jiaqian Ren, Hao Peng, Lei Jiang, Zhiwei Liu, Jia Wu, Zhengtao Yu, Philip S. Yu
Class Imbalance Boundary Aware Social Event Detection Class Uncertainty

September 14, 2023

Flexible Visual Recognition by Evidential Modeling of Confusion and Ignorance
Lei Fan, Bo Liu, Haoxiang Li, Ying Wu, Gang Hua
Visual Recognition Prediction Uncertainty Evidential Model Class Uncertainty Cell Phone Recognition

July 3, 2023

Robust Uncertainty Estimation for Classification of Maritime Objects
Jonathan Becktor, Frederik Scholler, Evangelos Boukas, Lazaros Nalpantidis
Classification Code Uncertainty Estimation Reaction Ship Uncertainty Measure Robust Uncertainty Class Uncertainty Ship Dataset

June 1, 2023

The Risks of Recourse in Binary Classification
Hidde Fokkema, Damien Garreau, Tim van Erven
Simple Classifier Binary Classification Risk Sensitive Machine Learning System Algorithmic Recourse Robust Recourse Class Uncertainty

May 5, 2023

Reduction of Class Activation Uncertainty with Background Information
H M Dipu Kabir
Vision Transformer Multitask Learning Class Activation Constructive Reduction Background Knowledge Improved Generalization High Performing Class Uncertainty

March 3, 2023

Uncertainty Estimation by Fisher Information-based Evidential Deep Learning
Danruo Deng, Guangyong Chen, Yang Yu, Furui Liu, Pheng-Ann Heng
Uncertainty Estimation Evidential Deep Learning Fisher Information Fisher Weighted Evidential Neural Network Class Uncertainty Uncertainty Estimation Task

December 5, 2022

Minimum Class Confusion based Transfer for Land Cover Segmentation in Rural and Urban Regions
Metehan Yalçın, Ahmet Alp Kındıroğlu, Furkan Burak Bağcı, Ufuk Uyan, Mahiye Uluyağmur Öztürk
Semantic Segmentation Transfer Learning Segmentation Performance Formality Transfer Urban Environment Rural Connectivity Class Uncertainty Land Cover Segmentation Satellite Image Segmentation

December 1, 2022

Test-Time Mixup Augmentation for Data and Class-Specific Uncertainty Estimation in Deep Learning Image Classification
Hansang Lee, Haeil Lee, Helen Hong, Junmo Kim
Raw Data Image Classification Uncertainty Estimation Mixup Augmentation Class Uncertainty