Data-Free Knowledge Distillation for Deep Neural Networks

Abstract

Recent advances in model compression have provided procedures for compressing large neural networks to a fraction of their original size while retaining most if not all of their accuracy. However, all of these approaches rely on access to the original training set, which might not always be possible if the network to be compressed was trained on a very large dataset, or on a dataset whose release poses privacy or safety concerns as may be the case for biometrics tasks. We present a method for data-free knowledge distillation, which is able to compress deep neural networks trained on large-scale datasets to a fraction of their size leveraging only some extra metadata to be provided with a pretrained model release. We also explore different kinds of metadata that can be used with our method, and discuss tradeoffs involved in using each of them.

Accepted to NIPS 2017 Workshop on Learning with Limited Data. Under review at AISTATS 2018

References in corpus (4)

• Distilling the Knowledge in a Neural Network
• FitNets: Hints for Thin Deep Nets
• Discrete Signal Processing on Graphs
• YouTube-8M: A Large-Scale Video Classification Benchmark

Cited by in corpus (9)

• Data-Free Adversarial Distillation
• Relational Knowledge Distillation
• Dream Distillation: A Data-Independent Model Compression Framework
• Adaptive Neural Network-Based Approximation to Accelerate Eulerian Fluid Simulation
• MSD: Multi-Self-Distillation Learning via Multi-classifiers within Deep Neural Networks
• EdgeAI: A Vision for Deep Learning in IoT Era
• DeGAN : Data-Enriching GAN for Retrieving Representative Samples from a Trained Classifier
• New Directions in Distributed Deep Learning: Bringing the Network at Forefront of IoT Design
• Impression Space from Deep Template Network