A Unified Plug-and-Play Framework for Effective Data Denoising and Robust Abstention

Abstract

The success of Deep Neural Networks (DNNs) highly depends on data quality. Moreover, predictive uncertainty makes high performing DNNs risky for real-world deployment. In this paper, we aim to address these two issues by proposing a unified filtering framework leveraging underlying data density, that can effectively denoise training data as well as avoid predicting uncertain test data points. Our proposed framework leverages underlying data distribution to differentiate between noise and clean data samples without requiring any modification to existing DNN architectures or loss functions. Extensive experiments on multiple image classification datasets and multiple CNN architectures demonstrate that our simple yet effective framework can outperform the state-of-the-art techniques in denoising training data and abstaining uncertain test data.

Under review

References in corpus (6)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks
• A Closer Look at Memorization in Deep Networks
• WebVision Database: Visual Learning and Understanding from Web Data
• Combating Label Noise in Deep Learning Using Abstention

Cited by in corpus (3)

• Controlled abstention neural networks for identifying skillful predictions for regression problems
• Controlled abstention neural networks for identifying skillful predictions for classification problems
• Reducing Risk and Uncertainty of Deep Neural Networks on Diagnosing COVID-19 Infection