A curated list of resources for Learning with Feature Noise
-
2016-NC - Noise detection in the Meta-Learning Level. [Paper] [Additional information]
-
2018-Arxiv - An Analysis of Active Learning With Uniform Feature Noise. [Paper]
-
2019-Soft Computing - A taxonomy on impact of label noise and feature noise using machine learning techniques. [Paper]
-
2021-ICLR - Learning with Feature-Dependent Label Noise: A Progressive Approach. [Paper]
-
2023-Arxiv - Freeze then Train: Towards Provable Representation Learning under Spurious Correlations and Feature Noise. [Paper]
-
2023-Arxiv - LCEN: A Novel Feature Selection Algorithm for Nonlinear, Interpretable Machine Learning Models. [Paper]
-
2023-IEEE - Feature Noise Boosts DNN Generalization under Label Noise. [Paper]
-
2023-JMLR - The Power of Contrast for Feature Learning: A Theoretical Analysis. [Paper]
-
2024-IJCNN - Feature-Aware Noise Contrastive Learning For Unsupervised Red Panda Re-Identification. [Paper]
-
2017-NIPS - Mixup: Beyond Empirical Risk Minimization. [Paper] [Code]
-
2019-ICML - Manifold Mixup: Better Representations by Interpolating Hidden States. [Paper] [Code]
-
2019-ICCV - CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features. [Paper] [Code]
-
2020-ICML - Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal Mixup. [Paper] [Code]
-
2020-NeurIPS - Boundary Thickness and Robustness in Learning Models. [Paper] [Code]
-
2016-CVPR - Rethinking the Inception Architecture for Computer Vision. [Paper]
-
2017-Arxiv - Towards Deep Learning Models Resistant to Adversarial Attacks. [Paper] [Code]
- 2019-Soft Computing - A taxonomy on impact of label noise and feature noise using machine learning techniques. [Paper]
The challenges of feature noise and label noise in machine learning are intricately linked. Many methods that address one type of noise often have implications for the other. For instance, techniques developed for handling noisy labels can sometimes be adapted to manage noisy features, and vice versa. Addressing both types of noise is crucial for developing robust machine learning models that perform well on real-world data.
Feel free to contribute to this list by adding relevant papers and resources. If you have suggestions or new papers to add, please send a pull request to this repository.
- Advances-in-Label-Noise-Learning
- Awesome-Noisy-Labels
- Search 'Noisy Label' Results
- Noisy Labels with Jupyter Notebook
- Noisy Label Neural Network1-Tensorflow
- Noisy Label Neural Network2-Chainer
- Multi-tasking Learning With Unreliable Labels
- Keras-noisy-lables-finetune
- Light CNN for Deep Face Recognition, in Tensorflow
- Rankpruning
- Cleanlab: machine learning python package for learning with noisy labels and finding label errors in datasets
- Deep Learning with Label Noise
- Deep Learning for Segmentation When Experts Disagree with Each Other
- Fair classification with group label noise
This list includes contributions from various researchers and practitioners in the field of machine learning. Special thanks to the Awesome-Learning-with-Label-Noise repository for providing valuable resources and inspiration. Contributions from the broader community are always welcome to keep this list up-to-date and comprehensive. If you have suggestions or new papers to add, please send a pull request to this repository.
By collaborating and sharing knowledge, we can advance our understanding and develop more robust solutions to the challenges posed by feature and label noise in machine learning.