<ul><li>Performance gap across classes remains a persistent challenge in machine learning, often attributed to variations in class hardness.</li><li>Hardness-based resampling is a promising approach to mitigate performance disparities.</li><li>Resampling does not meaningfully affect class-wise performance disparities, contrary to theoretical expectations.</li><li>Detailed analyses help identify key challenges unique to hardness-based imbalance and provide guidelines for future research.</li></ul>

Identifying Key Challenges of Hardness-Based Resampling

Discover more