Search for: All records

Abstract Machine unlearning is a cutting‐edge technology that embodies the privacy legal principle of the right to be forgotten within the realm of machine learning (ML). It aims to remove specific data or knowledge from trained models without retraining from scratch and has gained significant attention in the field of artificial intelligence in recent years. However, the development of machine unlearning research is associated with inherent vulnerabilities and threats, posing significant challenges for researchers and practitioners. In this article, we provide the first comprehensive survey of security and privacy issues associated with machine unlearning by providing a systematic classification across different levels and criteria. Specifically, we begin by investigating unlearning‐based security attacks, where adversaries exploit vulnerabilities in the unlearning process to compromise the security of machine learning (ML) models. We then conduct a thorough examination of privacy risks associated with the adoption of machine unlearning. Additionally, we explore existing countermeasures and mitigation strategies designed to protect models from malicious unlearning‐based attacks targeting both security and privacy. Further, we provide a detailed comparison between machine unlearning‐based security and privacy attacks and traditional malicious attacks. Finally, we discuss promising future research directions for security and privacy issues posed by machine unlearning, offering insights into potential solutions and advancements in this evolving field. 

Causality—the principle stating that the output of a system cannot temporally precede the input—is a universal property of nature. Here, we show that analogous input-output relations can also be realized in the spectral domain by leveraging the peculiar properties of time-modulated non-Hermitian photonic systems. Specifically, we uncover the existence of a broad class of complex time-modulated metamaterials that obey the time-domain equivalent of the well-established frequency-domain Kramers–Kronig relations (a direct consequence of causality). We find that, in the scattering response of such time-modulated systems, the output frequencies are inherently prohibited from spectrally preceding the input frequencies, and hence we refer to these systems as “spectrally causal.” We explore the consequences of this newly introduced concept for several relevant applications, including broadband perfect absorption, temporal cloaking of an “event,” and truly unidirectional propagation along a synthetic dimension. By emulating the concept of causality in the spectral domain and providing new tools to extend the field of temporally modulated metamaterials (“chrono-metamaterials”) into the complex realm, our findings may open unexplored opportunities and enable relevant technological advances in various areas of photonics and, more broadly, of wave physics and engineering.