Instance Cross Entropy for Deep Metric Learning
This work addresses the challenge of improving deep metric learning for applications like image retrieval and recognition, though it appears incremental as it builds on existing cross entropy concepts.
The authors tackled the problem of designing effective loss functions for deep metric learning by proposing instance cross entropy (ICE), which measures the difference between estimated and ground-truth instance-level matching distributions, and demonstrated its superiority with extensive experiments on three real-world benchmarks.
Loss functions play a crucial role in deep metric learning thus a variety of them have been proposed. Some supervise the learning process by pairwise or tripletwise similarity constraints while others take advantage of structured similarity information among multiple data points. In this work, we approach deep metric learning from a novel perspective. We propose instance cross entropy (ICE) which measures the difference between an estimated instance-level matching distribution and its ground-truth one. ICE has three main appealing properties. Firstly, similar to categorical cross entropy (CCE), ICE has clear probabilistic interpretation and exploits structured semantic similarity information for learning supervision. Secondly, ICE is scalable to infinite training data as it learns on mini-batches iteratively and is independent of the training set size. Thirdly, motivated by our relative weight analysis, seamless sample reweighting is incorporated. It rescales samples' gradients to control the differentiation degree over training examples instead of truncating them by sample mining. In addition to its simplicity and intuitiveness, extensive experiments on three real-world benchmarks demonstrate the superiority of ICE.