R2-D2: Repetitive Reprediction Deep Decipher for Semi-Supervised Deep Learning
This provides a principled framework for semi-supervised learning, addressing a foundational issue in deep learning with incremental improvements.
The paper tackled the lack of theoretical justification for using network predictions as pseudo-labels in semi-supervised deep learning, proposing the R2-D2 method that outperformed state-of-the-art methods by 5 percentage points on ImageNet.
Most recent semi-supervised deep learning (deep SSL) methods used a similar paradigm: use network predictions to update pseudo-labels and use pseudo-labels to update network parameters iteratively. However, they lack theoretical support and cannot explain why predictions are good candidates for pseudo-labels in the deep learning paradigm. In this paper, we propose a principled end-to-end framework named deep decipher (D2) for SSL. Within the D2 framework, we prove that pseudo-labels are related to network predictions by an exponential link function, which gives a theoretical support for using predictions as pseudo-labels. Furthermore, we demonstrate that updating pseudo-labels by network predictions will make them uncertain. To mitigate this problem, we propose a training strategy called repetitive reprediction (R2). Finally, the proposed R2-D2 method is tested on the large-scale ImageNet dataset and outperforms state-of-the-art methods by 5 percentage points.