Paper Info
Title
Knowledge Distillation Via Instance Relationship Graph
Abstract
The key challenge of knowledge distillation is to extract general,moderate and sufficient knowledge from a teacher network to guide a student network. In this paper a novel Instance Relationship Graph (IRG) is proposed for knowledge distillation. It models three kinds of knowledge, including instance features, instance relationships and feature space transformation, while the latter two kinds of knowledge are neglected by previous methods. Firstly, the IRG is constructed to model the distilled knowledge of one network layer by considering instance features and instance relationships as vertexes and edges respectively. Secondly, an IRG transformation is proposed to models the feature space transformation across layers. It is more moderate than directly mimicking the features at intermediate layers. Finally, hint loss functions are designed to force a student's IRGs to mimic the structures of a teacher's IRGs. The proposed method effectively captures the knowledge along the whole network via IRGs, and thus shows stable convergence and strong robustness to different network architectures. In addition, the proposed method shows superior performance over existing methods on datasets of various scales.
Year
DOI
Venue
2019
10.1109/CVPR.2019.00726
2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019)
Field
DocType
ISSN
Graph,Pattern recognition,Computer science,Theoretical computer science,Distillation,Artificial intelligence
Conference
1063-6919
Citations 
PageRank 
References 
13
0.53
0
Authors
7
Name
Order
Citations
PageRank
Yufan Liu1153.93
Jiajiong Cao2141.21
Li Bing3542.75
Chunfeng Yuan441830.84
Weiming Hu55300261.38
Li Yangxi6345.75
Yunqiang Duan7130.53