phuber
by dmizr
[Re] Can gradient clipping mitigate label noise? (ML Reproducibility Challenge 2020)
AI summary
Gradient clipping mitigation
An implementation of gradient clipping as a method to mitigate the effects of noisy labels in machine learning models
- stars
- 14
- forks
- 6
- watching
- 3
Similar projects
Found by comparing what the projects do, not just their names.
Label noise mitigation method
An implementation of a PyTorch-based deep learning method to improve robustness against noisy labels in image classification tasks
Label noise mitigation
Develops and evaluates machine learning algorithms to mitigate the effects of noisy labels in supervised learning.
Gradient smoothing
PyTorch implementation of a technique to improve the interpretability of deep learning models by adding noise to the gradients
Label noise correction
A method to train deep learning classifiers on noisy labels using a small set of trusted data
Label correction library
Provides PyTorch implementation of a method to address noisy labels in medical image segmentation.
Noisy label solver
Develops a robust learning framework to handle noisy labels in multimodal data and improve cross-modal retrieval.
Semi-supervised learner
A PyTorch implementation of a semi-supervised learning framework for training deep neural networks with noisy labels by dynamically dividing the data into clean and noisy sets.
Label noise mitigator
An implementation of a method to improve model robustness against inherent label noise in machine learning models
Noisy label solver
A PyTorch implementation of a method for learning with noisy labels in deep neural networks
Deep learning on noisy data
An implementation of training deep neural networks on noisy labels with bootstrapping using Keras
Label correction mechanism
This project explores how to adapt neural networks to noisy labels by introducing a mechanism that can learn to correct the errors.
Federated Learning Reverser
An algorithm that breaks secure aggregation protocols in federated learning by recovering individual model updates from aggregated sums
Gradient inversion vulnerability fix
An implementation of a method to invert gradients in federated learning to potentially reveal sensitive client data
Model optimizer
Enables the creation of smaller neural network models through efficient pruning and quantization techniques
Label noise correction algorithm
An implementation of an unsupervised label noise modeling and loss correction approach for deep learning.
