@inproceedings{Zhang-FSE-20,
author = {Zhang, Ziqi and Li, Yuanchun and Guo, Yao and Chen, Xiangqun and
Liu, Yunxin},
title = {Dynamic Slicing for Deep Neural Networks},
year = {2020},
isbn = {9781450370431},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3368089.3409676},
doi = {10.1145/3368089.3409676},
abstract = {Program slicing has been widely applied in a variety of software
engineering tasks. However, existing program slicing techniques only deal
with traditional programs that are constructed with instructions and
variables, rather than neural networks that are composed of neurons and
synapses. In this paper, we introduce NNSlicer, the first approach for
slicing deep neural networks based on data-flow analysis. Our method
understands the reaction of each neuron to an input based on the difference
between its behavior activated by the input and the average behavior over
the whole dataset. Then we quantify the neuron contributions to the slicing
criterion by recursively backtracking from the output neurons, and calculate
the slice as the neurons and the synapses with larger contributions. We
demonstrate the usefulness and effectiveness of NNSlicer with three
applications, including adversarial input detection, model pruning, and
selective model protection. In all applications, NNSlicer significantly
outperforms other baselines that do not rely on data flow analysis.},
booktitle = {Proceedings of the 28th ACM Joint Meeting on European Software
Engineering Conference and Symposium on the Foundations of Software
Engineering},
pages = {838¨C850},
numpages = {13},
keywords = {Deep neural networks, data-flow analysis, dynamic slicing,
program slicing},
location = {Virtual Event, USA},
series = {ESEC/FSE 2020}
}