Deepfakes have rapidly emerged as a serious threat to society due to their ease of creation and dissemination, triggering the accelerated development of detection technologies. However, many existing detectors rely on lab-generated datasets for validation, which may not prepare them for novel, real-world deepfakes. This paper extensively reviews and analyzes state-of-the-art deepfake detectors, evaluating them against several critical criteria. These criteria categorize detectors into 4 high-level groups and 13 fine-grained sub-groups, aligned with a unified conceptual framework we propose. This classification offers practical insights into the factors affecting detector efficacy. We evaluate the generalizability of 16 leading detectors across comprehensive attack scenarios, including black-box, white-box, and gray-box settings. Our systematized analysis and experiments provide a deeper understanding of deepfake detectors and their generalizability, paving the way for future research and the development of more proactive defenses against deepfakes.

The proliferation of deepfake technologies poses urgent challenges and serious risks to digital integrity, particularly within critical sectors such as forensics, journalism, and the legal system. While existing detection systems have made significant progress in classification accuracy, they typically function as black-box models, offering limited transparency and minimal support for human reasoning. This hinders their usability in real-world decision-making contexts, especially for non-expert users. We present DF-P2E (Deepfake: Prediction to Explanation), a novel multimodal framework that integrates visual, semantic, and narrative layers of explanation to make deepfake detection interpretable and accessible. The framework consists of three modular components: (1) a deepfake classifier with Grad-CAM-based saliency visualisation, (2) a visual captioning module that generates natural language summaries of manipulated regions, and (3) a narrative refinement module that uses a fine-tuned LLM to produce context-aware, user-sensitive explanations. We instantiate and evaluate the framework on the DF40 benchmark, the most diverse deepfake dataset to date. Experiments demonstrate that our system achieves competitive detection performance while providing high-quality explanations aligned with Grad-CAM activations. By unifying prediction and explanation in a coherent, human-aligned pipeline, this work offers a scalable approach to interpretable deepfake detection, advancing the broader vision of trustworthy and transparent AI systems for media forensics.

Deepfake detection remains a pressing challenge, particularly in real-world settings where smartphone-captured media from digital screens often introduces Moiré artifacts that can distort detection outcomes. This study systematically evaluates state-of-the-art (SOTA) deepfake detectors on Moiré-affected videos—an issue that has received little attention. We collected a dataset of 12,832 videos, spanning 35.64 hours, from Celeb-DF, DFD, DFDC, UADFV, and FF++ datasets, capturing footage under diverse real-world conditions, including varying screens, smartphones, lighting setups, and camera angles. To further examine the influence of Moiré patterns on deepfake detection, we conducted additional experiments using our DeepMoiréFake, referred to as (DMF) dataset, and two synthetic Moiré generation techniques. Across 15 top-performing detectors, our results show that Moiré artifacts degrade performance by as much as 25.4\%, while synthetically generated Moiré patterns lead to a 21.4\% drop in accuracy. Surprisingly, demoiréing methods, intended as a mitigation approach, instead worsened the problem, reducing accuracy by up to 16\%. These findings underscore the urgent need for detection models that can robustly handle Moiré distortions alongside other real-world challenges, such as compression, sharpening, and blurring. By introducing the DMF dataset, we aim to drive future research toward closing the gap between controlled experiments and practical deepfake detection.

The increasing complexity of decision-making in dynamic environments, particularly in high-stakes domains like cybersecurity, demands more than automated solutions—it requires effective integration of human expertise with advanced AI capabilities. While approaches like ensemble learning and Mixture of Experts (MoE) enhance automated decision-making, they struggle with handling uncertainty and novel scenarios. Techniques such as learning to defer and learning to complement mitigate this by incorporating human input, but assume that a definitive expert is always available—an assumption that often fails in real-world settings. To bridge this gap, we introduce A 2 C , a modular, multi-stage collaborative decision-making framework that enhances adaptability and decision robustness under uncertainty by seamlessly transitioning between three decision-making modes: Automated, Augmented Deferral, and Collaborative Exploration (CoEx). A key innovation of CoEX is its ability to handle cases where both AI and human experts face uncertainty, overcoming a critical limitation of traditional deferral systems. We validate A 2 C through experiments on benchmark datasets, Large Language Model (LLM) simulations of human–AI collaboration, and real-world human–AI interaction studies with cybersecurity researchers. Results show that A 2 C consistently outperforms conventional approaches that rely solely on full automation or selective human intervention, demonstrating its potential as a practical and scalable solution for expert decision-making in complex domains. For image detection on CIFAR-10, detection rates improved from 37.8% with automation alone to 64.75% with augmented deferral, and further to 92.95% with collaborative exploration. Similarly, for intrusion detection on KDDCup, rates rose from 33.43% with automation to 35.18% with augmented deferral, and finally reached 87.04% with CoEx, highlighting its effectiveness in handling uncertainty.

Deepfake detection is critical in mitigating the societal threats posed by manipulated videos. While various algorithms have been developed for this purpose, challenges arise when detectors operate externally, such as on smartphones, when users take a photo of deepfake images and upload on the Internet. One significant challenge in such scenarios is the presence of Moiré patterns, which degrade image quality and confound conventional classification algorithms, including deep neural networks (DNNs). The impact of Moiré patterns remains largely unexplored for deepfake detectors. In this study, we investigate how camera-captured deepfake videos from digital screens affect detector performance. We conducted experiments using two prominent datasets, CelebDF and FF++, comparing the performance of four state-of-theart detectors on camera-captured deepfake videos with introduced Moiré patterns. Our findings reveal a significant decline in detector accuracy, with none achieving above 68% on average. This underscores the critical need to address Moiré pattern challenges in real-world deepfake detection scenarios.

Security Operations Centres (SOCs) play a pivotal role in defending organisations against evolving cyber threats. They function as central hubs for detecting, analysing, and responding promptly to cyber incidents with the primary objective of ensuring the confidentiality, integrity, and availability of digital assets. However, they struggle against the growing problem of alert fatigue, where the sheer volume of alerts overwhelms SOC analysts and raises the risk of overlooking critical threats. In recent times, there has been a growing call for human-AI teaming, wherein humans and AI collaborate with each other, leveraging their complementary strengths and compensating for their weaknesses. The rapid advances in AI and the growing integration of AI-enabled tools and technologies within SOCs give rise to a compelling argument for the implementation of human-AI teaming within the SOC environment. Therefore, in this article, we present our vision for human-AI teaming to address the problem of alert fatigue in the SOC. We propose the 𝒜 2 𝒞 Framework, which enables flexible and dynamic decision making by allowing seamless transitions between automated, augmented, and collaborative modes of operation. Our framework allows AI-powered automation for routine alerts, AI-driven augmentation for expedited expert decision making, and collaborative exploration for tackling complex, novel threats. By implementing and operationalising 𝒜 2 𝒞, SOCs can significantly reduce alert fatigue while empowering analysts to efficiently and effectively respond to security incidents.

In recent times, there has been a notable surge in the amount of vision and sensing/time-series data obtained from drones and satellites.This data can be utilized in various fields, such as precision agriculture, disaster management, environmental monitoring, and others.However, the analysis of such data poses significant challenges due to its complexity, heterogeneity, and scale.Furthermore, it is critical to identify anomalies and maintain/monitor the health of drones and satellite systems to enable the aforementioned applications and sciences.This workshop presents an excellent opportunity to explore solutions that specifically target the detection of anomalies and novel occurrences in drones and satellite systems and their data.For more information, visit our website at https://sites.google.com/view/ansd23.

The development of techniques for creating completely synthetic photographic images and videos, as well as the increasing prevalence of disinformation associated with such synthetic media, have sparked interest in the computer community. One such issue is the proliferation of deepfakes. However, much of the research is devoted to outwitting the SOTA in order to generate and detect fabricated images and video. Considering them from the perspective of computer security and human ethics is largely ignored. As such, this workshop aims to provide a forum for researchers to exchange ideas and methodologies, as well as to provide academia and industry with insights into the development and identification of fake media from a computer security perspective.

Time series anomaly detection is extensively studied in statistics, economics, and computer science. Over the years, numerous methods have been proposed for time series anomaly detection using deep learning-based methods. Many of these methods demonstrate state-of-the-art performance on benchmark datasets, giving the false impression that these systems are robust and deployable in many practical and industrial real-world scenarios. In this paper, we demonstrate that the performance of state-of-the-art anomaly detection methods is degraded substantially by adding only small adversarial perturbations to the sensor data. We use different scoring metrics such as prediction errors, anomaly, and classification scores over several public and private datasets ranging from aerospace applications, server machines, to cyber-physical systems in power plants. Under well-known adversarial attacks from Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD) methods, we demonstrate that state-of-the-art deep neural networks (DNNs) and graph neural networks (GNNs) methods, which claim to be robust against anomalies and have been possibly integrated in real-life systems, have their performance drop to as low as 0%. To the best of our understanding, we demonstrate, for the first time, the vulnerabilities of anomaly detection systems against adversarial attacks. The overarching goal of this research is to raise awareness towards the adversarial vulnerabilities of time series anomaly detectors.

Recent advancements in web-based multimedia technologies, such as face recognition web services powered by deep learning, have been significant. As a result, companies such as Microsoft, Amazon, and Naver provide highly accurate commercial face recognition web services for a variety of multimedia applications. Naturally, such technologies face persistent threats, as virtually anyone with access to deepfakes can quickly launch impersonation attacks. These attacks pose a serious threat to authentication services, which rely heavily on the performance of their underlying face recognition technologies. Despite its gravity, deepfake abuse involving commercial web services and their robustness have not been thoroughly measured and investigated. By conducting a case study on celebrity face recognition, we examine the robustness of black-box commercial face recognition web APIs and open-source tools against Deepfake Impersonation (DI) attacks. While the majority of APIs do not make specific claims of deepfake robustness, we find that authentication mechanisms may get built one top of them, nonetheless. We demonstrate the vulnerability of face recognition technologies to DI attacks, achieving respective success rates of 78.0% for targeted (TA) attacks; we also propose mitigation strategies, lowering respective attack success rates to as low as 1.26% for TA attacks with adversarial training.

Over the last few decades, artificial intelligence research has made tremendous strides, but it still heavily relies on fixed datasets in stationary environments. Continual learning is a growing field of research that examines how AI systems can learn sequentially from a continuous stream of linked data in the same way that biological systems do. Simultaneously, fake media such as deepfakes and synthetic face images have emerged as significant to current multimedia technologies. Recently, numerous method has been proposed which can detect deepfakes with high accuracy. However, they suffer significantly due to their reliance on fixed datasets in limited evaluation settings. Therefore, in this work, we apply continuous learning to neural networks' learning dynamics, emphasizing its potential to increase data efficiency significantly. We propose Continual Representation using Distillation (CoReD) method that employs the concept of Continual Learning (CL), Representation Learning (RL), and Knowledge Distillation (KD). We design CoReD to perform sequential domain adaptation tasks on new deepfake and GAN-generated synthetic face datasets, while effectively minimizing the catastrophic forgetting in a teacher-student model setting. Our extensive experimental results demonstrate that our method is efficient at domain adaptation to detect low-quality deepfakes videos and GAN-generated images from several datasets, outperforming the-state-of-art baseline methods.

Deep learning-based video manipulation methods have become widely accessible to the masses. With little to no effort, people can quickly learn how to generate deepfake (DF) videos. While deep learning-based detection methods have been proposed to identify specific types of DFs, their performance suffers for other types of deepfake methods, including real-world deepfakes, on which they are not sufficiently trained. In other words, most of the proposed deep learning-based detection methods lack transferability and generalizability. Beyond detecting a single type of DF from benchmark deepfake datasets, we focus on developing a generalized approach to detect multiple types of DFs, including deepfakes from unknown generation methods such as DeepFake-in-the-Wild (DFW) videos. To better cope with unknown and unseen deepfakes, we introduce a Convolutional LSTM-based Residual Network (CLRNet), which adopts a unique model training strategy and explores spatial as well as the temporal information in a deepfakes. Through extensive experiments, we show that existing defense methods are not ready for real-world deployment. Whereas our defense method (CLRNet) achieves far better generalization when detecting various benchmark deepfake methods (97.57% on average). Furthermore, we evaluate our approach with a high-quality DeepFake-in-the-Wild dataset, collected from the Internet containing numerous videos and having more than 150,000 frames. Our CLRNet model demonstrated that it generalizes well against high-quality DFW videos by achieving 93.86% detection accuracy, outperforming existing state-of-the-art defense methods by a considerable margin.

Reducing false positives while detecting anomalies is of growing importance for various industrial applications and mission-critical infrastructures, including satellite systems. Undesired false positives can be costly for such systems, bringing the operation to a halt for human experts to determine if the anomalies are true anomalies that need to be mitigated. Although rule-based or machine learning-based anomaly detection approaches have been studied, a tensor-based decomposition method has not been extensively explored. In this work, we introduce an Integrative Tensor-based Anomaly Detection (ITAD) framework to detect anomalies in a satellite system with the goal of minimizing false positives. We construct 3rd-order tensors with telemetry data collected from the Korea Multi-Purpose Satellite-2 (KOMPSAT-2) and calculate the anomaly score using one of the component matrices obtained by applying CANDECOMP/PARAFAC decomposition to detect anomalies. Our result shows that our tensor-based approach outperforms existing methods, achieving higher accuracy and lower false positive rates. And we successfully deployed our anomaly detection system in real KOMPSAT-2 mission operation.

Detecting an anomaly is not only important for many terrestrial applications on Earth but also for space applications. Especially, satellite missions are highly risky because unexpected hardware and software failures can occur due to sudden or unforeseen space environment changes. Anomaly detection and spacecraft health monitoring systems have heavily relied on human expertise to investigate whether they are a true anomaly or not. Also, it is practically infeasible to produce labels on data due to the enormous amount of telemetries generated from a satellite. In this work, we propose a data-driven anomaly detection algorithm for Korea Multi-Purpose Satellite 2 (KOMPSAT-2). We develop a Multivariate Convolution LSTM with Mixtures of Probabilistic Principal Component Analyzers, where our approach uses both neural networks and probabilistic clustering to improve the anomaly detection performance. We evaluated our approach with a total of 22 million telemetry samples collected for 10 months from KOMPSAT-2. We also compare our approach with other state-of-the-art approaches. We show that our proposed approach is 35.8% better in precision, and 18.2% better in F-1 score than the best baseline approach. We plan to deploy our algorithm in the second half of 2019 to actually apply real operation of KOMPSAT-2.

While GPS (Global Positioning System) navigation and GPS based autonomous car driving techniques are mature, the security of GPS signal has not been a primary system concern. In fact, it has been shown that an ordinary GPS can be easily spoofed by a low-cost, open-source based software defined radio (SDR) system such as bladeRF and HackRF [3, 4] which can cause serious complications to the navigation system of the car especially for self-driving cars which are driven based on the information from sensors, cameras, and GPS. In this work, we design a new GPS authentication system called lightweight authentication GPS (AuthGPS) to authenticate GPS signal against GPS spoofing and LTE base station broadcast message spoofing. In the past, there have been many successful attempts on GPS spoofing which resulted in shifting the destination of the car to the spoofer's desired location. In a case where the car is connected to the Internet via LTE network, the navigation system can find out if the GPS is spoofed by acquiring the correct GPS satellite information from LTE base stations' location information. However, a skillful attacker can spoof the LTE base station signals [7] as well using another SDR which will produce spoofed LTE base station information. Hence giving wrong information about GPSinformation according to the will of the attacker. Currently, there are defense methods against the GPS spoofing [6]. One of them is signal-processing-based methods. By monitoring unusual or unreasonable signal changes at GPS receivers, GPS spoofing attack can be detected. Received Power Monitoring (RPM) looks at all the received amplitude and automatic gain control (AGC) setpoint [1]. The receiver will sense drastic power jump if a spoofing attack occurs. However, an overly powerful spoofing attack with noise is not detectable in the case of this mechanism. Another way to detect spoofing is the symmetric-key encryption mechanism of GPS signals. According to previous research [2], encrypted precision code for anti-spoofing referred to as P(Y) code, might not be able to be spoofed. A spoofing attack can be detected by calculating cross-correlation between spoofed coarse/acquisition code, referred to as C/A code, and P(Y) code. Unfortunately, this P(Y) code is only for military purpose while the C/A code can be publicly accessible. This method requires knowledge of specific key which is not revealed to the public to decrypt P(Y) code. Monitoring the direction of arrival of the signals can be one of the methods [6]. A GPS receiver measures the direction-of-arrival vector with more than 3 antennas. Even though this signal-geometrybased system with multiple antennas makes GPS robust, an attacker can spoof signals from multiple directions, which is more difficult to detect the spoofing. The idea here is to develop a system that can discriminate the spoofed signal without complex computation or heavy message exchange. An important component of AuthGPS is a verification of valid GPS satellite information via LTE base stations' location information. We propose 6-digit one-time password-based authentication system in this paper.

We explored the problems which are soon to be faced while parking autonomous cars in parking lots. Like where is the closest parking slot available to the autonomous car? How to navigate to that location? What kind of parking structures could be good for autonomous cars? We also provide an initial solution which uses a central server and the graph of the parking lot to guide the cars to the closest parking slots. With experiments, we have shown that our proposed method should be effective for the controlled parking for self-driving cars.