Differentiable Top-k Classification Learning

15 Jun 2022  ·  Felix Petersen, Hilde Kuehne, Christian Borgelt, Oliver Deussen ·

The top-k classification accuracy is one of the core metrics in machine learning. Here, k is conventionally a positive integer, such as 1 or 5, leading to top-1 or top-5 training objectives. In this work, we relax this assumption and optimize the model for multiple k simultaneously instead of using a single k. Leveraging recent advances in differentiable sorting and ranking, we propose a differentiable top-k cross-entropy classification loss. This allows training the network while not only considering the top-1 prediction, but also, e.g., the top-2 and top-5 predictions. We evaluate the proposed loss function for fine-tuning on state-of-the-art architectures, as well as for training from scratch. We find that relaxing k does not only produce better top-5 accuracies, but also leads to top-1 accuracy improvements. When fine-tuning publicly available ImageNet models, we achieve a new state-of-the-art for these models.

PDF Abstract


Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Image Classification ImageNet Top-k DiffSortNets (EfficientNet-L2) Top 1 Accuracy 88.37% # 58


No methods listed for this paper. Add relevant methods here