EXACT: How to Train Your Accuracy
Not on the map. Why?
Abstract (arXiv)
Classification tasks are usually evaluated in terms of accuracy. However, accuracy is discontinuous and cannot be directly optimized using gradient ascent. Popular methods minimize cross-entropy, hinge loss, or other surrogate losses, which can lead to suboptimal results. In this paper, we propose a new optimization framework by introducing stochasticity to a model's output and optimizing expected accuracy, i.e. accuracy of the stochastic model. Extensive experiments on linear models and deep image classification show that the proposed optimization method is a powerful alternative to widely used classification losses.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 25 of its 57 references that are in the map. First posted on arXiv 19 May 2022.
Builds on
25 of 57 references.
References · 25
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2014Adam: A Method for Stochastic Optimizationcited by 36,960
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
- 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
- 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
- 2013Auto-Encoding Variational Bayescited by 3,980
- 2016Wide Residual Networkscited by 3,057
- 2015Character-level Convolutional Networks for Text Classificationcited by 2,293
- 2020Knowledge Distillation: A Surveycited by 1,578
- 2019TabNet: Attentive Interpretable Tabular Learningcited by 640
- 2017Generative Adversarial Networks: An Overviewcited by 588
- 2021ASAM: Adaptive Sharpness-Aware Minimization for Scale-Invariant Learning of Deep Neural Networkscited by 293
13 more
- 2019Probabilistic Face Embeddingscited by 171
- 2016Deep Pyramidal Residual Networkscited by 163
- 2020Data Uncertainty Learning in Face Recognitioncited by 134
- 2020Bayesian Neural Networks: An Introduction and Surveycited by 130
- 2020Large-scale Robust Deep AUC Maximization: A New Surrogate Loss and Empirical Studies on Medical Image Classificationcited by 84
- 2021Recall@k Surrogate Loss with Large Batches and Similarity Mixupcited by 49
- 2020Calibrated Surrogate Losses for Adversarially Robust Classificationcited by 36
- 2020An Ensemble of Simple Convolutional Neural Network Models for MNIST Digit Recognitioncited by 30
- 2021sigmoidF1: A Smooth F1 Score Surrogate Loss for Multilabel Classificationcited by 23
- 2019Learning Surrogate Lossescited by 21
- 2018Accelerated Training for Massive Classification via Dynamic Class Selectioncited by 16
- 2022Probabilistic Embeddings Revisitedcited by 8
- 2015A New Smooth Approximation to the Zero One Loss with a Probabilistic Interpretationcited by 2
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.