Deep Metric Learning via Facility Location
This work addresses a key bottleneck in deep metric learning for applications like image retrieval and clustering, offering a novel structured prediction framework that enhances global embedding optimization.
The paper tackles the problem of performance degradation in deep metric learning due to local training unaware of global structure, proposing a global metric learning scheme that achieves state-of-the-art results on clustering and retrieval tasks across multiple datasets, with improvements in NMI and Recall@K metrics.
Learning the representation and the similarity metric in an end-to-end fashion with deep networks have demonstrated outstanding results for clustering and retrieval. However, these recent approaches still suffer from the performance degradation stemming from the local metric training procedure which is unaware of the global structure of the embedding space. We propose a global metric learning scheme for optimizing the deep metric embedding with the learnable clustering function and the clustering metric (NMI) in a novel structured prediction framework. Our experiments on CUB200-2011, Cars196, and Stanford online products datasets show state of the art performance both on the clustering and retrieval tasks measured in the NMI and Recall@K evaluation metrics.