Class Prior

Class priors, representing the prior probabilities of different classes in a dataset, are crucial in many machine learning tasks, particularly classification. Current research focuses on mitigating the limitations of relying on known or assumed class priors, exploring techniques like contrastive learning and adapting classifiers to handle class imbalances and shifts in real-world deployments. This research is significant because accurate class prior estimation or handling their absence improves model performance, fairness, and robustness, especially in scenarios with limited labeled data, noisy labels, or evolving data distributions. The impact spans various applications, including image recognition, power quality event classification, and remote sensing.

Papers

September 7, 2024

Fine-Grained Representation Learning via Multi-Level Contrastive Learning without Class Priors
Houwang Jiang, Zhuxian Liu, Guodong Liu, Xiaolong Liu, Shihua Zhan
Contrastive Learning Fine Grained Contrastive Loss Unsupervised Visual Representation Learning COntrastive Disentanglement Class Prior

April 28, 2024

Naive Bayes Classifiers and One-hot Encoding of Categorical Variables
Christopher K. I. Williams
Categorical Data Naive Bayes Classifier Posterior Probability Class Prior One Hot Encoding Probability Vector

February 22, 2024

Enhancing Power Quality Event Classification with AI Transformer Models
Ahmad Mohammad Saber, Amr Youssef, Davor Svetinovic, Hatem Zeineldin, Deepa Kundur, Ehab El-Saadany
Deep Learning Deep Learning Framework Class Prior Power Quality Voltage Imaging

January 25, 2024

Class-attribute Priors: Adapting Optimization to Heterogeneity and Fairness Objective
Xuechen Zhang, Mingchen Li, Jiasi Chen, Christos Thrampoulidis, Samet Oymak
Loss Function Optimization Purpose Optimal Classifier Fairness Objective Unknown Heterogeneity Class Prior Label Imbalance

December 27, 2023

Twice Class Bias Correction for Imbalanced Semi-Supervised Learning
Lan Li, Bowen Tao, Lu Han, De-chuan Zhan, Han-jia Ye
Semi Supervised Learning Class Imbalance Class Distribution Bias Correction Imbalanced Semi Supervised Learning Class Prior

September 25, 2023

Can Class-Priors Help Single-Positive Multi-Label Learning?
Biao Liu, Ning Xu, Jie Wang, Xin Geng
Multi Label Classification Class Prior Single Positive Multi Label

September 10, 2023

Mitigating Word Bias in Zero-shot Prompt-based Classifiers
Adian Liusie, Potsawee Manakul, Mark J. F. Gales
Language Model Zero Shot Classification Prompt Based Language Bias Zero Shot Prompt Class Prior Threshold Optimization

September 4, 2023

Adapting Classifiers To Changing Class Priors During Deployment
Natnael Daba, Bruce McIntosh, Abhijit Mahalanobis
Supervised Learning Simple Classifier Software Deployment Learning Generalizable Class Prior Adaptive Classification

August 29, 2023

Class Prior-Free Positive-Unlabeled Learning with Taylor Variational Loss for Hyperspectral Remote Sensing Imagery
Hengwei Zhao, Xinyu Wang, Jingtao Li, Yanfei Zhong
LeArning Abstract Natural Gradient Hyperspectral Image Classification Class Prior Pu Learning Hyperspectral Remote Sensing Variational Loss

June 12, 2023

Making Binary Classification from Multiple Unlabeled Datasets Almost Free of Supervision
Yuhao Wu, Xiaobo Xia, Jun Yu, Bo Han, Gang Niu, Masashi Sugiyama, Tongliang Liu
Binary Classification Class Prior

August 7, 2022

Class-Incremental Learning with Cross-Space Clustering and Controlled Transfer
Arjun Ashok, K J Joseph, Vineeth Balasubramanian
Incremental Learning Class Incremental Learning Formality Transfer Class Prior

June 10, 2022

Lightweight Conditional Model Extrapolation for Streaming Data under Class-Prior Shift
Paulina Tomaszewska, Christoph H. Lampert
Non Stationary Multi Class Classification Class Distribution Streaming Data Trend Extrapolation Class Prior Adaptive Classification

June 6, 2022

Class Prior Estimation under Covariate Shift: No Problem?
Dirk Tasche
Covariate Shift Right Problem Relevant Covariates Target Distribution Class Prior Prior Value Estimate