Improving Deep Metric Learning with Virtual Classes and Examples Mining
This work addresses a specific bottleneck in deep metric learning for researchers and practitioners, offering an incremental improvement over existing generation-based methods.
The paper tackles the problem of sampling hard negative examples in deep metric learning by introducing MIRAGE, a method that uses virtual classes composed of generated examples as buffer areas between training classes, resulting in significant improvements on datasets like Cub-200-2011, Cars-196, and Stanford Online Products compared to other generation methods.
In deep metric learning, the training procedure relies on sampling informative tuples. However, as the training procedure progresses, it becomes nearly impossible to sample relevant hard negative examples without proper mining strategies or generation-based methods. Recent work on hard negative generation have shown great promises to solve the mining problem. However, this generation process is difficult to tune and often leads to incorrectly labelled examples. To tackle this issue, we introduce MIRAGE, a generation-based method that relies on virtual classes entirely composed of generated examples that act as buffer areas between the training classes. We empirically show that virtual classes significantly improve the results on popular datasets (Cub-200-2011, Cars-196 and Stanford Online Products) compared to other generation methods.