<ul data-eligibleForWebStory="false"><li>Researchers propose a new undersampling approach to tackle imbalanced data classification issues by avoiding synthetic data pitfalls and under-fitting.</li><li>Their method selects datapoints based on their potential to improve model loss rather than randomly undersampling majority data.</li><li>The approach aims to identify an optimal subset of majority training data by rejecting redundant datapoints, leveraging a bilevel optimization problem.</li><li>Experimental results demonstrate F1 scores up to 10% higher compared to existing state-of-the-art methods.</li></ul>

A Bilevel Optimization Framework for Imbalanced Data Classification

Discover more