Type
Conference PaperAuthors
Hamdi, Abdullah
Ghanem, Bernard

KAUST Department
Electrical Engineering ProgramPhysical Science and Engineering (PSE) Division
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
KAUST Grant Number
OSR-CRG2018-3730Date
2021-01-10Preprint Posting Date
2019-04-09Online Publication Date
2021-01-10Print Publication Date
2020Permanent link to this record
http://hdl.handle.net/10754/660663
Metadata
Show full item recordAbstract
Despite the impressive performance of Deep Neural Networks (DNNs) on various vision tasks, they still exhibit erroneous high sensitivity toward semantic primitives (e.g. object pose). We propose a theoretically grounded analysis for DNN robustness in the semantic space. We qualitatively analyze different DNNs’ semantic robustness by visualizing the DNN global behavior as semantic maps and observe interesting behavior of some DNNs. Since generating these semantic maps does not scale well with the dimensionality of the semantic space, we develop a bottom-up approach to detect robust regions of DNNs. To achieve this, we formalize the problem of finding robust semantic regions of the network as optimizing integral bounds and we develop expressions for update directions of the region bounds. We use our developed formulations to quantitatively evaluate the semantic robustness of different popular network architectures. We show through extensive experimentation that several networks, while trained on the same dataset and enjoying comparable accuracy, do not necessarily perform similarly in semantic robustness. For example, InceptionV3 is more accurate despite being less semantically robust than ResNet50. We hope that this tool will serve as a milestone towards understanding the semantic robustness of DNNs.Citation
Hamdi, A., & Ghanem, B. (2020). Towards Analyzing Semantic Robustness of Deep Neural Networks. Lecture Notes in Computer Science, 22–38. doi:10.1007/978-3-030-66415-2_2Sponsors
This work was supported by the King Abdullah University of Science and Technology (KAUST) Office of Sponsored Research under Award No. OSR-CRG2018-3730Publisher
Springer NatureConference/Event name
European Conference on Computer Vision (ECCV) 2020ISBN
97830306641459783030664152
arXiv
1904.04621Additional Links
http://link.springer.com/10.1007/978-3-030-66415-2_2Relations
Is Supplemented By:- [Software]
Title: ajhamdi/semantic-robustness: implementation of the paper: "Towards Analyzing Semantic Robustness of Deep Neural Networks" (ECCV 2020 workshop). Publication Date: 2019-03-20. github: ajhamdi/semantic-robustness Handle: 10754/667397
ae974a485f413a2113503eed53cd6c53
10.1007/978-3-030-66415-2_2