Active poisoning: efficient backdoor attacks on transfer learning-based brain-computer interfaces

Jiang, Xue; Meng, Lubin; Li, Siyang; Wu, Dongrui

doi:10.1007/s11432-022-3548-2

Cited by 8 publications

(2 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When this poisoned data is used in TL training, the backdoor is successfully set. As a result, it reached an average of around 90% attack success rate using various model architectures and datasets [71]. Another backdoor attack model was proposed by Shuo Wang et al, in which the goal was to defeat defenses such as pruning, retraining, and input preprocessing.…”

Section: Attacks On Transfer Learningmentioning

confidence: 99%

Distributed Learning in the IoT–Edge–Cloud Continuum

Arzovs,

Judvaitis,

Nesenbergs

et al. 2024

MAKE

View full text Add to dashboard Cite

The goal of the IoT–Edge–Cloud Continuum approach is to distribute computation and data loads across multiple types of devices taking advantage of the different strengths of each, such as proximity to the data source, data access, or computing power, while mitigating potential weaknesses. Most current machine learning operations are currently concentrated on remote high-performance computing devices, such as the cloud, which leads to challenges related to latency, privacy, and other inefficiencies. Distributed learning approaches can address these issues by enabling the distribution of machine learning operations throughout the IoT–Edge–Cloud Continuum by incorporating Edge and even IoT layers into machine learning operations more directly. Approaches like transfer learning could help to transfer the knowledge from more performant IoT–Edge–Cloud Continuum layers to more resource-constrained devices, e.g., IoT. The implementation of these methods in machine learning operations, including the related data handling security and privacy approaches, is challenging and actively being researched. In this article the distributed learning and transfer learning domains are researched, focusing on security, robustness, and privacy aspects, and their potential usage in the IoT–Edge–Cloud Continuum, including research on tools to use for implementing these methods. To achieve this, we have reviewed 145 sources and described the relevant methods as well as their relevant attack vectors and provided suggestions on mitigation.

show abstract

Section: Attacks On Transfer Learningmentioning

confidence: 99%

Distributed Learning in the IoT–Edge–Cloud Continuum

Arzovs,

Judvaitis,

Nesenbergs

et al. 2024

MAKE

View full text Add to dashboard Cite

show abstract

“…These attacks exploit the model's sensitivity to minor perturbations in input data. Jiang et al, (2023) explored and understood the potential risks associated with backdoor attacksaboutTL for EEG-based BCIs. This involves assessing the vulnerabilities that can be introduced when transferring models trained on poisoned data.…”

Section: Backdoor and Adversarial Attacks On Image Security Using Tra...mentioning

confidence: 99%

A Review Study on various Image Security Techniques and Emerging trends for Visual Data Protection

Dr. K. Rajeswari, Mrs. Dipali Ghatge

2023

tjjpt

View full text Add to dashboard Cite

In the rapidly evolving digital age, the proliferation of visual data and images has led to increased concerns about their security and protection. This paper reviews various image security techniques and emerging trends aimed at safeguarding visual data in a digital landscape. The research explores the evolution of image security measures and their adaptation to address contemporary challenges in data protection. This study introduces the stage by elucidating the multifaceted nature of image security, emphasizing its significance in preserving the confidentiality and integrity of visual data. The complex landscape of image security threats is examined, highlighting the challenges faced in safeguarding visual data from various vulnerabilities. This paper delves into the need and significance of image security in the digital age, presenting a compelling case for the adoption of robust security measures. It explores the risks associated with image theft and tampering in the online world and underscores the essential role of image encryption and authentication methods in countering these threats. The study then delves into the realm of encryption techniques, offering insights into both symmetric and asymmetric cryptographic algorithms designed to enhance image security on the internet. The research extends its focus to secure data hiding techniques, such as reversible data hiding in encrypted images, histogram shifting-based data hiding, and high-embedding capacity methods employing optimization algorithms. The paper also explores digital watermarking as a means of image authentication and copyright protection, along with image tamper detection and self-recovery mechanisms in watermarking schemes. Additionally, this review assesses the impact of emerging trends, including machine learning, Blockchain, and AI-based solutions, on the security of visual data, such as deep neural networks, generative adversarial neural networks, and their role in image steganography, forensics, and privacy protection. Consequently, by providing an in-depth analysis of the current state of image security and future directions, this study contributes to the understanding of the critical role played by image security in the digital age and its implications for data protection and privacy.

show abstract