<ul data-eligibleForWebStory="true"><li>Feature shifts between data sources are common in various applications, causing issues like erroneous features.</li><li>Localizing shifted features is crucial to correct or filter data and maintain downstream analysis integrity.</li><li>Detecting distribution shifts is feasible, but localizing the originating features remains a challenge.</li><li>Existing solutions for localizing feature shifts are either inaccurate or not scalable for large datasets.</li><li>A new approach, the Feature Shift Localization Network (FSL-Net), is introduced in this work.</li><li>FSL-Net is a neural network designed to quickly and accurately localize feature shifts in large and high-dimensional datasets.</li><li>The network is trained with diverse datasets to learn statistical properties and can identify shifts in new datasets without re-training.</li><li>The FSL-Net model and code are available for public use on GitHub at https://github.com/AI-sandbox/FSL-Net.</li></ul>

Feature Shift Localization Network

Discover more