Vibration-based Uncertainty Estimation for Learning from Limited Supervision
We investigate the problem of estimating uncertainty for training data, so that deep neural networks can make use of the results for learning from limited supervision. However, neither the prediction probability nor the entropy can accurately capture the uncertainty of out-of-distribution data. In this paper, we present a novel approach that measures the uncertainty from the vibration of sequential data, \textit{e.g.}, the output probability during the training procedure. The key observation is that, a training sample that suffers heavier vibration often offers richer information when it is manually labeled. We make use of the Fourier Transformation to measure the extent of vibration, deriving a powerful tool that can be used for semi-supervised, active learning, and one-bit supervision. Experiments on the CIFAR10, CIFAR100, and mini-ImageNet datasets validate the effectiveness of our approach.
PDF Abstract