Balanced Meta-Softmax for Long-Tailed Visual Recognition
By Jiawei Ren, Cunjun Yu, Shunan Sheng, Xiao Ma, Haiyu Zhao, Shuai Yi, Hongsheng Li. 2020
801 citations recorded by the source corpus.
Deep classifiers have achieved great success in visual recognition. However, real-world data is long-tailed by nature, leading to the mismatch between training and testing distributions. In this paper, we show that Softmax function, though used in most classification tasks, gives a biased gradient estimation under the long-tailed setup. This paper presents Balanced Softmax, an elegant unbiased extension of Softmax, to accommodate the label distribution shift between training and testing. Theoretically, we derive the generalization bound for multiclass Softmax regression and show our loss minimizes the bound. In addition, we introduce Balanced Meta-Softmax, applying a complementary Meta Sampler to estimate the optimal class sample rate and further improve long-tailed learning. In our experiments, we demonstrate that Balanced Meta-Softmax outperforms state-of-the-art long-tailed classification solutions on both visual recognition and instance segmentation tasks.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Meta & Continual Learning
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
In Computer Vision and Pattern Recognition
In Meta & Continual Learning
Citations in the corpus
- Reviving Iterative Training with Mask Guidance for Interactive Segmentation (2021, neutral): “These datasets stimulated the development of state-of-the-art instance segmentation algorithms [42, 43, 44, 45], providing a large variety of labeled classes with a sufficient number of examples.”
- Balanced Activation for Long-tailed Visual Recognition (2020, neutral): “A future research direction could be improving the optimization process in complementary to Balanced Activation, for example, a learnable importance sampling method in [8].”
- Parametric Contrastive Learning (2021, methodology): “For CIFAR-100-LT, we strictly follow the setting of [38] for fair comparison.”
- Disentangling Label Distribution for Long-tailed Visual Recognition (2020, methodology): “We compare our LADE with PC Softmax and the current state-of-the-art methods, Causal Norm [54], and Balanced Softmax [46].”
- Balanced Knowledge Distillation for Long-tailed Learning (2021, background): “The key to this line of work [18, 29, 7, 43, 55, 39] is to adjust learning focus by modifying objective functions.”