Graph-Based Fuzz Testing for Deep Learning Inference Engine
August 13, 2020 ยท Declared Dead ยท ๐ International Conference on Software Engineering
"No code URL or promise found in abstract"
Evidence collected by the PWNC Scanner
Authors
Weisi Luo, Dong Chai, Xiaoyue Run, Jiang Wang, Chunrong Fang, Zhenyu Chen
arXiv ID
2008.05933
Category
cs.SE: Software Engineering
Citations
42
Venue
International Conference on Software Engineering
Last Checked
3 months ago
Abstract
With the wide use of Deep Learning (DL) systems, academy and industry begin to pay attention to their quality. Testing is one of the major methods of quality assurance. However, existing testing techniques focus on the quality of DL models but lacks attention to the core underlying inference engines (i.e., frameworks and libraries). Inspired by the success stories of fuzz testing, we design a graph-based fuzz testing method to improve the quality of DL inference engines. This method is naturally followed by the graph structure of DL models. A novel operator-level coverage criterion based on graph theory is introduced and six different mutations are implemented to generate diversified DL models by exploring combinations of model structures, parameters, and data inputs. The Monte Carlo Tree Search (MCTS) is used to drive DL model generation without a training process. The experimental results show that the MCTS outperforms the random method in boosting operator-level coverage and detecting exceptions. Our method has discovered more than 40 different exceptions in three types of undesired behaviors: model conversion failure, inference failure, output comparison failure. The mutation strategies are useful to generate new valid test inputs, by up to 8.2% more operator-level coverage on average and 8.6 more exceptions captured.
Community Contributions
Found the code? Know the venue? Think something is wrong? Let us know!
๐ Similar Papers
In the same crypt โ Software Engineering
R.I.P.
๐ป
Ghosted
R.I.P.
๐ป
Ghosted
GraphCodeBERT: Pre-training Code Representations with Data Flow
R.I.P.
๐ป
Ghosted
DeepTest: Automated Testing of Deep-Neural-Network-driven Autonomous Cars
R.I.P.
๐ป
Ghosted
Microservices: yesterday, today, and tomorrow
R.I.P.
๐ป
Ghosted
Devign: Effective Vulnerability Identification by Learning Comprehensive Program Semantics via Graph Neural Networks
R.I.P.
๐ป
Ghosted
A Survey of Machine Learning for Big Code and Naturalness
Died the same way โ ๐ป Ghosted
R.I.P.
๐ป
Ghosted
Language Models are Few-Shot Learners
R.I.P.
๐ป
Ghosted
PyTorch: An Imperative Style, High-Performance Deep Learning Library
R.I.P.
๐ป
Ghosted
XGBoost: A Scalable Tree Boosting System
R.I.P.
๐ป
Ghosted