Learning from Label Proportions with Instance-wise Consistency
Not on the map. Why?
Abstract (arXiv)
Learning from Label Proportions (LLP) is a weakly supervised learning method that aims to perform instance classification from training data consisting of pairs of bags containing multiple instances and the class label proportions within the bags. Previous studies on multiclass LLP can be divided into two categories according to the learning task: per-instance label classification and per-bag label proportion estimation. However, these methods often results in high variance estimates of the risk when applied to complex models, or lack statistical learning theory arguments. To address this issue, we propose new learning methods based on statistical learning theory for both per-instance and per-bag policies. We demonstrate that the proposed methods are respectively risk-consistent and classifier-consistent in an instance-wise manner, and analyze the estimation error bounds. Additionally, we present a heuristic approximation method that utilizes an existing method for regressing label proportions to reduce the computational complexity of the proposed methods. Through benchmark experiments, we demonstrated the effectiveness of the proposed methods.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 33 of its 72 references that are in the map. First posted on arXiv 24 Mar 2022.
Builds on
33 of 72 references.
References · 33
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2014Adam: A Method for Stochastic Optimizationcited by 37,240
- 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,349
- 2018Attention-based Deep Multiple Instance Learningcited by 1,095
- 2018Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labelscited by 1,059
- 2016Temporal Ensembling for Semi-Supervised Learningcited by 1,057
- 2016Making Deep Neural Networks Robust to Label Noise: a Loss Correction Approachcited by 665
- 2018Deep Learning for Classical Japanese Literaturecited by 422
- 2017Size-Independent Sample Complexity of Neural Networkscited by 290
- 2017Positive-Unlabeled Learning with Non-Negative Risk Estimatorcited by 219
- 2016A vector-contraction inequality for Rademacher complexitiescited by 142
- 2020Progressive Identification of True Labels for Partial-Label Learningcited by 102
21 more
- 2017Learning with Biased Complementary Labelscited by 93
- 2020Provably Consistent Partial-Label Learningcited by 92
- 2017Learning from Complementary Labelscited by 88
- 2018On the Minimal Supervision for Training Any Binary Classifier from Only Unlabeled Datacited by 43
- 2014On Learning from Label Proportionscited by 38
- 2018Classification from Pairwise Similarity and Unlabeled Datacited by 36
- 2019Learning from Label Proportions with Consistency Regularizationcited by 34
- 2019Deep multi-class learning from label proportionscited by 33
- 2022Learning from Label Proportions by Learning with Label Noisecited by 30
- 2019Mitigating Overfitting in Supervised Classification from Two Unlabeled Datasets: A Consistent Risk Correction Approachcited by 29
- 2020Unbiased Risk Estimators Can Mislead: A Case Study of Learning with Complementary Labelscited by 29
- 2017Co-training for Demographic Classification Using Deep Learning from Label Proportionscited by 28
- 2019Learning from Label Proportions with Generative Adversarial Networkscited by 25
- 2017Binary Classification from Positive-Confidence Datacited by 23
- 2017Weakly Supervised Classification in High Energy Physicscited by 21
- 2019Classification from Pairwise Similarities/Dissimilarities and Unlabeled Data via Empirical Risk Minimizationcited by 21
- 2021Two-stage Training for Learning from Label Proportionscited by 19
- 2018Deep Learning from Label Proportions for Emphysema Quantificationcited by 13
- 2021Learning with Proper Partial Labelscited by 8
- 2018Label Propagation for Learning with Label Proportionscited by 5
- 2021Fast learning from label proportions with small bagscited by 5
Built on by
Citing papers per year: 2023: 1; 2024: 1.
2 papers.