KDCTime: Knowledge Distillation with Calibration on InceptionTime for Time-series Classification
This work addresses the few-shot problem in time-series classification for researchers and practitioners, offering an incremental improvement over existing methods.
The paper tackles overfitting in time-series classification on UCR datasets by proposing KDCTime, a method that uses knowledge distillation with calibration on InceptionTime, achieving promising accuracy while reducing inference time by two orders of magnitude compared to ROCKET.
Time-series classification approaches based on deep neural networks are easy to be overfitting on UCR datasets, which is caused by the few-shot problem of those datasets. Therefore, in order to alleviate the overfitting phenomenon for further improving the accuracy, we first propose Label Smoothing for InceptionTime (LSTime), which adopts the information of soft labels compared to just hard labels. Next, instead of manually adjusting soft labels by LSTime, Knowledge Distillation for InceptionTime (KDTime) is proposed in order to automatically generate soft labels by the teacher model. At last, in order to rectify the incorrect predicted soft labels from the teacher model, Knowledge Distillation with Calibration for InceptionTime (KDCTime) is proposed, where it contains two optional calibrating strategies, i.e. KDC by Translating (KDCT) and KDC by Reordering (KDCR). The experimental results show that the accuracy of KDCTime is promising, while its inference time is two orders of magnitude faster than ROCKET with an acceptable training time overhead.