本文是計算機視覺領域頂級會議ICCV2021入選論文《An Empirical Study of the Collapsing Problem in Semi-Supervised 2D Human Pose Estimation(半監督二維人體姿態估計中的模型坍塌問題研究)》的解讀。該論文由北京大學王亦洲課題組與微軟亞洲研究院合作完成,研究了一致性學習算法在二維人體姿態估計中的模型坍塌問題,并提出了構建預測難度存在差異的樣本對來解決該問題。本文提出的算法可以有效地利用無標記數據中的信息,顯著提高了姿態估計模型的泛化能力。
[1] Kihyuk Sohn, David Berthelot, Chun-Liang Li, Zizhao Zhang, Nicholas Carlini, Ekin D Cubuk, Alex Kurakin, Han Zhang, and Colin Raffel. Fixmatch: Simplifying semisupervised learning with consistency and confidence. In Advances in Neural Information Processing Systems, 2020.
[2] David Berthelot, Nicholas Carlini, Ian Goodfellow, Nicolas Papernot, Avital Oliver, and Colin A Raffel. Mixmatch: A holistic approach to semi-supervised learning. In Advances in Neural Information Processing Systems, pages 5049–5059, 2019.