论文标题
通过路由多样化的分销专家，长尾认可
Long-tailed Recognition by Routing Diverse Distribution-Aware Experts
论文作者
论文摘要
自然数据通常是在语义类别上分布长尾的。现有的识别方法通过将更多的重点放在尾巴数据上，通过重新平衡/重新平衡/重新加权或在不同的数据组上结合，从而解决了这种不平衡的分类，从而导致尾巴准确性提高，但头部精度降低。 我们对培训数据进行动态视图，并随着训练数据的波动提供原则的模型偏见和方差分析：现有的长尾分类器总是会增加模型差异，并且由于对尾巴的艰难消极质量的混乱越来越多，因此头尾模型偏置差距仍然很大。 我们提出了一个新的长尾分类器，称为Routing Diverse Diverss（Ride）。它可以减少多个专家的模型差异，减少模型偏差，并通过分配感知的多样性损失，通过动态专家路由模块降低计算成本。在CIFAR100-LT，ImagEnet-LT和Inaturalist 2018基准上，Ride的最终表现优于最先进的5％至7％。它也是一个通用框架，适用于各种骨干网络，长尾算法和培训机制，可持续增长。我们的代码可在以下网址提供：https：//github.com/frank-xwang/ride-longtailrecognition。
Natural data are often long-tail distributed over semantic classes. Existing recognition methods tackle this imbalanced classification by placing more emphasis on the tail data, through class re-balancing/re-weighting or ensembling over different data groups, resulting in increased tail accuracies but reduced head accuracies. We take a dynamic view of the training data and provide a principled model bias and variance analysis as the training data fluctuates: Existing long-tail classifiers invariably increase the model variance and the head-tail model bias gap remains large, due to more and larger confusion with hard negatives for the tail. We propose a new long-tailed classifier called RoutIng Diverse Experts (RIDE). It reduces the model variance with multiple experts, reduces the model bias with a distribution-aware diversity loss, reduces the computational cost with a dynamic expert routing module. RIDE outperforms the state-of-the-art by 5% to 7% on CIFAR100-LT, ImageNet-LT and iNaturalist 2018 benchmarks. It is also a universal framework that is applicable to various backbone networks, long-tailed algorithms, and training mechanisms for consistent performance gains. Our code is available at: https://github.com/frank-xwang/RIDE-LongTailRecognition.
