Exact Reformulation and Optimization for Direct Metric Optimization in Binary Imbalanced Classification
Not on the map. Why?
Abstract (arXiv)
For classification with imbalanced class frequencies, i.e., imbalanced classification (IC), standard accuracy is known to be misleading as a performance measure. While most existing methods for IC resort to optimizing balanced accuracy (i.e., the average of class-wise recalls), they fall short in scenarios where the significance of classes varies or certain metrics should reach prescribed levels. In this paper, we study two key classification metrics, precision and recall, under three practical binary IC settings: fix precision optimize recall (FPOR), fix recall optimize precision (FROP), and optimize F1-score (OFOS). Unlike existing methods that rely on smooth approximations to deal with the indicator function involved, we introduce, for the first time, exact constrained reformulations for these direct metric optimization (DMO) problems, which can be effectively solved by exact penalty methods. Experiment results on multiple benchmark datasets demonstrate the practical superiority of our approach over the state-of-the-art methods for the three DMO problems. We also expect our exact reformulation and optimization (ERO) framework to be applicable to a wide range of DMO problems for binary IC and beyond. Our code is available at https://github.com/sun-umn/DMO.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 48 of its 89 references that are in the map. First posted on arXiv 21 Jul 2025.
Builds on
48 of 89 references.
References · 48
- 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
- 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
- 2019CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparisoncited by 1,423
- 2019Learning Imbalanced Datasets with Label-Distribution-Aware Margin Losscited by 989
- 2019EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Taskscited by 807
- 2017Skin Lesion Analysis Toward Melanoma Detection: A Challenge at the 2017 International Symposium on Biomedical Imaging (ISBI), Hosted by the International Skin Imaging Collaboration (ISIC)cited by 549
- 2021Deep learning: a statistical viewpointcited by 240
- 2019Differentiation of Blackbox Combinatorial Solverscited by 226
- 2021Fit without fear: remarkable mathematical phenomena of deep learning through the prism of interpolationcited by 132
- 2020Tilted Empirical Risk Minimizationcited by 106
- 2022AUC Maximization in the Era of Big Data and AI: A Surveycited by 100
- 2020Smooth-AP: Smoothing the Path Towards Large-Scale Image Retrievalcited by 98
36 more
- 2021Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentationcited by 88
- 2018Combo Loss: Handling Input and Output Imbalance in Multi-Organ Segmentationcited by 76
- 2022A Survey on Long-Tailed Visual Recognitioncited by 75
- 2019Optimizing Rank-based Metrics with Blackbox Differentiationcited by 58
- 2018Two-Player Games for Efficient Non-Convex Constrained Optimizationcited by 55
- 2018Training Well-Generalizing Classifiers for Fairness Metrics and Other Data-Dependent Constraintscited by 53
- 2021Recall@k Surrogate Loss with Large Batches and Similarity Mixupcited by 49
- 2016Scalable Learning of Non-Decomposable Objectivescited by 41
- 2021Stochastic Optimization of Areas Under Precision-Recall Curves with Provable Convergencecited by 35
- 2021Analyzing Overfitting under Class Imbalance in Neural Networks for Image Segmentationcited by 34
- 2019SoDeep: a Sorting Deep net to learn ranking loss surrogatescited by 28
- 2024Boosted Conformal Prediction Intervalscited by 27
- 2020AP-Loss for Accurate One-Stage Object Detectioncited by 24
- 2021sigmoidF1: A Smooth F1 Score Surrogate Loss for Multilabel Classificationcited by 23
- 2023Complexity of Single Loop Algorithms for Nonlinear Programming with Stochastic Objective and Constraintscited by 16
- 2019GENO -- GENeric Optimization for Classical Machine Learningcited by 12
- 2024Selective Classification Under Distribution Shiftscited by 12
- 2024Variance-reduced first-order methods for deterministically constrained stochastic nonconvex optimization with strong convergence guaranteescited by 11
- 2018Optimizing Non-decomposable Measures with Deep Networkscited by 10
- 2021MetricOpt: Learning to Optimize Black-Box Evaluation Metricscited by 9
- 2021Implicit Rate-Constrained Optimization of Non-decomposable Objectivescited by 9
- 2019AP-Perf: Incorporating Generic Performance Metrics in Differentiable Learningcited by 8
- 2022Exploring the Algorithm-Dependent Generalization of AUPRC Optimization with List Stabilitycited by 8
- 2015Surrogate Functions for Maximizing Precision at the Topcited by 7
- 2022NCVX: A General-Purpose Optimization Solver for Constrained Machine and Deep Learningcited by 7
- 2025A Baseline Method for Removing Invisible Image Watermarks using Deep Image Priorcited by 7
- 2021The ensmallen library for flexible numerical optimizationcited by 6
- 2023Optimization and Optimizers for Adversarial Robustnesscited by 6
- 2020The Effect of Class Imbalance on Precision-Recall Curvescited by 5
- 2021A surrogate loss function for optimization of Fβ score in binary classification with imbalanced datacited by 5
- 2022Optimization for Classical Machine Learning Problems on the GPUcited by 5
- 2020Introduction to Nonsmooth Analysis and Optimizationcited by 4
- 2023Federated Learning with Convex Global and Local Constraintscited by 4
- 2024Pearls from Pebbles: Improved Confidence Functions for Auto-labelingcited by 3
- 2022Imbalanced Classification in Medical Imaging via Regroupingcited by 2
- 2025First-order methods for stochastic and finite-sum convex optimization with deterministic constraintscited by 1
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.