Poster

Abad, Gorka; Ersoy, Oğuzhan; Picek, Stjepan; Ramírez-Durán, Víctor Julio; Urbieta, Aitor

doi:10.1145/3548606.3563532

Cited by 5 publications

(20 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A multivariate dataset was used in this study. The author presented an FL-based [45] pedagogical analysis in the education sphere through a learning-based incentive mechanism. The author of this work offered an FL server that could be run locally without affecting the properties or data of the existing system.…”

Section: Literature Reviewmentioning

confidence: 99%

Secure IoMT for Disease Prediction Empowered With Transfer Learning in Healthcare 5.0, the Concept and Case Study

et al. 2023

View full text Add to dashboard Cite

Identifying human diseases remains a difficult process, even in the age of advanced information technology and the smart healthcare industry 5.0. In the smart healthcare industry 5.0, precise prediction of human diseases, particularly lethal cancer diseases, is critical for human well-being. The global Internet of Medical Things sector has advanced at a breakneck pace in recent years, from small wristwatches to large aircraft. The critical aspects of the Internet of Medical Things include security and privacy, owing to the massive scale and deployment of the Internet of Medical Things networks. Transfer learning with a secure IoMTbased approach is considered. The google net deep machine-learning model is used for accurate disease prediction in the smart healthcare industry 5.0. We can easily and reliably anticipate the lethal cancer disease in the human body by using the secure IoMT-based transfer learning approach. Furthermore, the results of the proposed secure IoMT-based Transfer learning techniques are used to validate the best cancer disease prediction in the smart healthcare industry 5.0. The proposed secure IoMT-based transfer learning methodology reached 98.8%, better than the state-of-the-art methodologies used previously for cancer disease prediction in the smart healthcare industry 5.0.INDEX TERMS IoMT, transfer learning, deep machine learning, histopathology, image processing, lung cancer.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Secure IoMT for Disease Prediction Empowered With Transfer Learning in Healthcare 5.0, the Concept and Case Study

et al. 2023

View full text Add to dashboard Cite

show abstract

“…Specifically, the attacker chooses a location, trigger size, and trigger polarity. Instead of evaluating these parameters exhaustively, we rely on prior research to select an already used (and well-performing) configuration [2]. Specifically, we employ a trigger size equivalent to 30% of the image, placing the trigger in the top-left corner with a polarity of 1 (light blue).…”

Section: Single Attackermentioning

confidence: 99%

“…This defense could be deployed by a device locally at inference time to reject inputs that have an abnormal entropy level, i.e., contain a trigger. Other methods, such as artificial brain stimulation (ABS) [23], are ineffective against SNNs with neuromorphic data [2]. STRIP considers access to a clean test set.…”

Section: Defense Evaluationmentioning

confidence: 99%

“…Although there is no precise method for measuring the stealthiness of a trigger in backdoor attacks or the perturbation in adversarial examples, mean squared error (MSE) and structural similarity index (SSIM) are commonly used. Following recent work [2], we also utilize these metrics for evaluating the stealthiness of the triggers. MSE compares the pixel-wise difference between the clean and the backdoored sample.…”

Section: Stealthinessmentioning

confidence: 99%

“…In the last few years, new threats have been discovered [17], such as adversarial examples, model inversion, and backdoor attacks, to name a few. Regarding SNNs, recent investigations have concluded that SNNs are also vulnerable to some of these attacks, i.e., adversarial examples [28,33,9] and backdoor attacks [2,1]. In the context of FL, security and privacy evaluations have also been in the scope of security experts [3,25], concluding that FL is vulnerable to privacy attacks, such as membership inference, and security attacks, such as backdoor attacks.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Sniper Backdoor: Single Client Targeted Backdoor Attack in Federated Learning

Abad

Paguada

Ersoy

et al. 2023

2023 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML)

Self Cite

View full text Add to dashboard Cite

This paper investigates the vulnerability of spiking neural networks (SNNs) and federated learning (FL) to backdoor attacks using neuromorphic data. Despite the efficiency of SNNs and the privacy advantages of FL, particularly in low-powered devices, we demonstrate that these systems are susceptible to such attacks. We first assess the viability of using FL with SNNs using neuromorphic data, showing its potential usage. Then, we evaluate the transferability of known FL attack methods to SNNs, finding that these lead to suboptimal attack performance. Therefore, we explore backdoor attacks involving single and multiple attackers to improve the attack performance. Our primary contribution is developing a novel attack strategy tailored to SNNs and FL, which distributes the backdoor trigger temporally and across malicious devices, enhancing the attack's effectiveness and stealthiness. In the best case, we achieve a 100% attack success rate, 0.13 MSE, and 98.9 SSIM. Moreover, we adapt and evaluate an existing defense against backdoor attacks, revealing its inadequacy in protecting SNNs. This study underscores the need for robust security measures in deploying SNNs and FL, particularly in the context of backdoor attacks.

show abstract