ICST 2024
Mon 27 - Fri 31 May 2024 Canada
Tue 28 May 2024 10:15 - 10:30 at Room 1 - Session 1 (Opening, Keynote, Papers 1)

Deep Learning (DL) models excel in computer vision tasks but can be susceptible to adversarial examples. This paper introduces Triple-Metric EvoAttack (TM-EVO), an efficient algorithm for evaluating the robustness of object-detection DL models against adversarial attacks. TM-EVO utilizes a multi-metric fitness function to guide an evolutionary search efficiently in creating effective adversarial test inputs with minimal perturbations. We evaluate TM-EVO on widely-used objectdetection DL models, DETR and Faster R-CNN, and open-source datasets, COCO and KITTI. Our findings reveal that TM-EVO outperforms the state-of-the-art EvoAttack baseline, leading to adversarial tests with less noise while maintaining efficiency.

Tue 28 May

Displayed time zone: Eastern Time (US & Canada) change

09:00 - 10:30
Session 1 (Opening, Keynote, Papers 1)AIST at Room 1
09:00
15m
Day opening
Workshop Opening
AIST
Gregory Gay Chalmers | University of Gothenburg, Sebastiano Panichella Zurich University of Applied Sciences, Aitor Arrieta Mondragon University
09:15
60m
Keynote
Towards Better Software Quality in the Era of Large Language Models
AIST
Lingming Zhang University of Illinois at Urbana-Champaign
10:15
15m
Talk
Generating Minimalist Adversarial Perturbations to Test Object-Detection Models: An Adaptive Multi-Metric Evolutionary Search Approach
AIST
Cristopher McIntyre-Garcia , Adrien Heymans University of Ottawa, Beril Borali University of Ottawa, Won-Sook Le University of Ottawa, Shiva Nejati University of Ottawa