입력 변이에 따른 딥러닝 모델 취약점 연구 및 검증
Analysis of Deep Learning Model Vulnerability According to Input Mutation
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
딥러닝 모델은 변이를 통해 훈련 데이터에서 벗어난 입력으로부터 잘못된 예측 결과를 산출할 수 있으며 이는 자율주행, 보안 분야 등에서 치명적인 사고로 이어질 수 있다. 딥러닝 모델의 신뢰성 보장을 위해서는 다양한 변이를통해 예외적인 상황에 대한 모델의 처리 능력이 검증되어야 한다. 하지만, 기존 연구가 제한된 모델을 대상으로만수행되었으며, 여러 입력 변이 유형에 구분을 짓지 않고 사용했다. 본 연구에서는 딥러닝 검증 데이터 세트로 널리사용되고 있는 CIFAR10 데이터 세트를 기반으로 다양한 상용화된 모델과 추가 버전을 포함하여 총 6개의 모델에대한 신뢰성 검증을 수행한다. 이를 위해 실생활에서 발생할 수 있는 6가지 유형의 입력 변이 알고리즘을 다양한 파라미터와 함께 데이터 세트에 개별적으로 적용하여 각각에 대한 모델의 정확도를 비교함으로써 특정 변이 유형과 관련된 모델의 취약점을 구체적으로 파악한다.
Abstract
The deep learning model can produce false prediction results due to inputs that deviate from training data throughvariation, which leads to fatal accidents in areas such as autonomous driving and security. To ensure reliability of themodel, the model's coping ability for exceptional situations should be verified through various mutations. However, previousstudies were carried out on limited scope of models and used several mutation types without separating them. Based on theCIFAR10 data set, widely used dataset for deep learning verification, this study carries out reliability verification for total ofsix models including various commercialized models and their additional versions. To this end, six types of input mutationalgorithms that may occur in real life are applied individually with their various parameters to the dataset to compare theaccuracy of the models for each of them to rigorously identify vulnerabilities of the models associated with a particularmutation type.
참고문헌 (12)
- He, Kaiming, "Deep residual learning for image recognition", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.770~778 (2016)
- Krizhevsky, Alex, "Imagenet classification with deep convolutional neural networks", Proceedings of the Advances in Neural Information Processing Systems (NIPS), 25, pp.1097~1105 (2012)
- Xiong, Wayne, "Achieving human parity in conversational speech recognition", arXiv preprint arXiv:1610.05256, ~ (2016)
- Julian, Kyle D., "Policy compression for aircraft collision avoidance systems", Proceedings of the IEEE, 2016, 1~10
- Bojarski, Mariusz, "End to end learning for self-driving cars", arXiv preprint arXiv:1604.07316, ~ (2016)
- Yuan, Zhenlong, "Droid-sec:deep learning in android malware detection", Proceedings of the ACM Conference on SIGCOMM, pp.371~372 (2014)
- Xie, Xiaofei, "Deephunter: A coverage-guided fuzz testing framework for deep neural networks", Proceedings of the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA), pp.146~157 (2019)
- Ma, Lei, "Deepmutation:Mutation testing of deep learning systems", Proceedings of the IEEE International Symposium on Software Reliability Engineering (ISSRE), pp.100~111 (2018)
- Tian, Yuchi, "Deeptest:Automated testing of deep-neural-network-driven autonomous cars", Proceedings of the International Conference on Software Engineering (ICSE), pp.303~314 (2018)
- "Who’s responsible when an autonomou s car crashes?" (2016)[URL]
- Madry, Aleksander, "Towards deep learning models resistant to adversarial attacks", arXiv preprint arXiv:1706.06083, ~ (2017)
- Meng, Dongyu, "Magnet: a two-pronged defense against adversarial examples", Proceedings of the ACM SIGSAC Conference on Computer and Communications Security (CCS), pp.135~147 (2017)