Robust and Resource-Efficient Data-Free Knowledge Distillation by Generative Pseudo Replay

by   Kuluhan Binici, et al.

Data-Free Knowledge Distillation (KD) allows knowledge transfer from a trained neural network (teacher) to a more compact one (student) in the absence of original training data. Existing works use a validation set to monitor the accuracy of the student over real data and report the highest performance throughout the entire process. However, validation data may not be available at distillation time either, making it infeasible to record the student snapshot that achieved the peak accuracy. Therefore, a practical data-free KD method should be robust and ideally provide monotonically increasing student accuracy during distillation. This is challenging because the student experiences knowledge degradation due to the distribution shift of the synthetic data. A straightforward approach to overcome this issue is to store and rehearse the generated samples periodically, which increases the memory footprint and creates privacy concerns. We propose to model the distribution of the previously observed synthetic samples with a generative network. In particular, we design a Variational Autoencoder (VAE) with a training objective that is customized to learn the synthetic data representations optimally. The student is rehearsed by the generative pseudo replay technique, with samples produced by the VAE. Hence knowledge degradation can be prevented without storing any samples. Experiments on image classification benchmarks show that our method optimizes the expected value of the distilled model accuracy while eliminating the large memory overhead incurred by the sample-storing methods.


page 1

page 2

page 3

page 4

page 6

page 7

page 8

page 9


Conditional Generative Data-Free Knowledge Distillation based on Attention Transfer

Knowledge distillation has made remarkable achievements in model compres...

Preventing Catastrophic Forgetting and Distribution Mismatch in Knowledge Distillation via Synthetic Data

With the increasing popularity of deep learning on edge devices, compres...

Data-Free Knowledge Distillation with Soft Targeted Transfer Set Synthesis

Knowledge distillation (KD) has proved to be an effective approach for d...

Beyond Classification: Knowledge Distillation using Multi-Object Impressions

Knowledge Distillation (KD) utilizes training data as a transfer set to ...

Momentum Adversarial Distillation: Handling Large Distribution Shifts in Data-Free Knowledge Distillation

Data-free Knowledge Distillation (DFKD) has attracted attention recently...

How to Teach: Learning Data-Free Knowledge Distillation from Curriculum

Data-free knowledge distillation (DFKD) aims at training lightweight stu...

Robustness and Diversity Seeking Data-Free Knowledge Distillation

Knowledge distillation (KD) has enabled remarkable progress in model com...