Long Short-Term Sample Distillation
arXiv:2003.00739
Abstract
In the past decade, there has been substantial progress at training increasingly deep neural networks. Recent advances within the teacher--student training paradigm have established that information about past training updates show promise as a source of guidance during subsequent training steps. Based on this notion, in this paper, we propose Long Short-Term Sample Distillation, a novel training policy that simultaneously leverages multiple phases of the previous training process to guide the later training updates to a neural network, while efficiently proceeding in just one single generation pass. With Long Short-Term Sample Distillation, the supervision signal for each sample is decomposed into two parts: a long-term signal and a short-term one. The long-term teacher draws on snapshots from several epochs ago in order to provide steadfast guidance and to guarantee teacher--student differences, while the short-term one yields more up-to-date cues with the goal of enabling higher-quality updates. Moreover, the teachers for each sample are unique, such that, overall, the model learns from a very diverse set of teachers. Comprehensive experimental results across a range of vision and NLP tasks demonstrate the effectiveness of this new training method.
arXiv commentpublished as a conference paper at AAAI 2020
References in corpus (7)
- Distilling the Knowledge in a Neural Network
- FitNets: Hints for Thin Deep Nets
- On Calibration of Modern Neural Networks
- Temporal Ensembling for Semi-Supervised Learning
- Do Deep Nets Really Need to be Deep?
- GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding
- Born Again Neural Networks