On the Security & Privacy in Federated Learning

Abstract

Advances in Machine Learning (ML) and its wide range of applications boosted its popularity. Recent privacy awareness initiatives as the EU General Data Protection Regulation (GDPR) -European Parliament and Council Regulation No 2016/679, subdued ML to privacy and security assessments. Federated Learning (FL) grants a privacy-driven, decentralized training scheme that improves ML models' security. The industry's fast-growing adaptation and security evaluations of FL technology exposed various vulnerabilities. Depending on the FL phase, i.e., training or inference, the adversarial actor capabilities, and the attack type threaten FL's confidentiality, integrity, or availability (CIA). Therefore, the researchers apply the knowledge from distinct domains as countermeasures, like cryptography and statistics.This work assesses the CIA of FL by reviewing the state-of-theart (SoTA) for creating a threat model that embraces the attack's surface, adversarial actors, capabilities, and goals. We propose the first unifying taxonomy for attacks and defenses by applying this model. Additionally, we provide critical insights extracted by applying the suggested novel taxonomies to the SoTA, yielding promising future research directions.

Federated Learning Based Privacy Ensured Sensor Communication in IoT Networks: A Taxonomy, Threats and Attacks

“…Integrity of training data is compromised by adversaries in data poisoning attacks and thus it reduces the overall accuracy of the model. Data poisoning attacks may happen in any one of these forms: clean-label or dirty-label [91], [92]. In cleanlabel attack, the adversary cannot change the labels of data in a dataset, but he places the poisoned data instance online and waits for the genuine client to label it and include it into their dataset.…”

“…In cleanlabel attack, the adversary cannot change the labels of data in a dataset, but he places the poisoned data instance online and waits for the genuine client to label it and include it into their dataset. This type of attack happens when the data is collected from the untrusted sources and are therefore very difficult to identify as the poisoned instance is labelled correctly [91]. However, in dirty-label attack, the adversary changes the label of genuine instances of the dataset into the targeted ones.…”

“…When the adversaries manipulate the local model parameters, the attack is known as gradient manipulation. Whereas, when the training rules are manipulated by the adversaries, the attack is known as training rule manipulation attack [22], [91]. Backdoor attacks are performed by adding the stealth backdoors to the global model such that the overall accuracy of the model is retained [93].…”

Security of federated learning for cloud‐edge intelligence collaborative computing

“…Integrity attack: It mainly refers to interfering with the prediction results of the model, making the output results of the model fail to meet the expected performance. 16 Attackers usually use targeted attack methods, such as backdoor attack and clean label attack.…”

Sniper Backdoor: Single Client Targeted Backdoor Attack in Federated Learning

“…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.…”