Artificial intelligence (AI) and machine learning (ML) have come a long way from the earlier days of conceptual theories, to being an integral part of today’s technological society. Rapid growth of AI/ML and their penetration within a plethora of civilian and military applications, while successful, has also opened new challenges and obstacles. With almost no human involvement required for some of the new decision-making AI/ML systems, there is now a pressing need to gain better insights into how these decisions are made. This has given rise to a new field of AI research, explainable AI (XAI). In this article, we present a survey of XAI characteristics and properties. We provide an indepth review of XAI themes, and describe the different methods for designing and developing XAI systems, both during and post model-development. We include a detailed taxonomy of XAI goals, methods, and evaluation, and sketch the major milestones in XAI research. An overview of XAI for security and cybersecurity of XAI systems is also provided. Open challenges are delineated, and measures for evaluating XAI system robustness are described.

Advances in artificial intelligence (AI) and wireless technology are driving forward the large deployment of interconnected smart technologies that constitute cyber–physical systems (CPSs) and Internet of Things (IoT) for many commercial and military applications. CPS is characterized by communication, computing, and control engineering based on a large volume of data originating from various devices, plants, sensors, etc. Wireless technologies have enabled the ease of networking and communications for both CPS and IoT, by providing massive and critical connectivity and control mechanisms. However, they are prone to challenges, such as low latency, throughput, and scheduling. Recent research trends focus on how to intelligently use data from CPS units to enhance wireless connectivity in CPS. AI tools, particularly AI systems and machine learning (ML) algorithms, have been widely applied in the literature to develop efficient schemes for wireless CPS/IoT. This article presents a review on the role of AI in wireless networking for CPS and IoT. In particular, we focus on ML paradigms, such as transfer learning (TL), distributed learning, and federated learning, that have evolved as building blocks for the utilization of large data for learning, adaptation, and predictions in CPS and IoT systems that leverage wireless networking. Furthermore, we also highlight challenges faced by current and future wireless networks pertaining to CPS/IoT, which are yet to be addressed.

Unmanned aerial vehicles (UAVs) have increasingly shown to be useful in civilian applications (such as agriculture, public safety, surveillance) and mission critical military applications. Despite the growth in popularity and applications, UAVs have also been used for malicious purposes. In such instances, their timely detection and identification has garnished rising interest from government, industry and academia. While much work has been done for detecting UAVs, there still exist limitations related to the impact of extreme environmental conditions and big dataset requirements. This paper proposes a novel ensemble deep learning framework that has hybrid synthetic and deep features to detect unauthorized or malicious UAVs by using acoustic, image/video and wireless radio frequency (RF) signals for robust UAV detection and classification. We present the performance evaluation of the proposed approach using numerical results obtained from experiments using acoustic, image/video and wireless RF signals. The proposed approach outperforms the existing related approaches for detecting malicious UAVs.

Causal inference is the idea of cause and effect; this fundamental area of sciences can be applied to problem space associated with Newton’s laws or the devastating COVID-19 pandemic. The cause explains the “why,” whereas the effect describes the “what.” The domain itself encompasses a plethora of disciplines from statistics and computer science to economics and philosophy. Recent advancements in machine learning and artificial intelligence systems have nourished a renewed interest in identifying and estimating the cause-and-effect relationship from the substantial amount of available observational data. This has resulted in various new studies aimed at providing novel methods for identifying and estimating causal inference. We include a detailed taxonomy of causal inference frameworks, methods, and evaluation. An overview of causality for security is also provided. Open challenges are detailed, and approaches for evaluating the robustness of causal inference methods are described. This article aims to provide a comprehensive survey on such studies of causality. We provide an in-depth review of causality frameworks and describe the different methods.

Direct oral anticoagulants (DOACs) targeting activated factor Xa (FXa) are used to prevent or treat thromboembolic disorders. DOACs reversibly bind to FXa and inhibit its enzymatic activity. However, DOAC treatment carries the risk of anticoagulant-associated bleeding. Currently, only one specific agent, andexanet alfa, is approved to reverse the anticoagulant effects of FXa-targeting DOACs (FXaDOACs) and control life-threatening bleeding. However, because of its mechanism of action, andexanet alfa requires a cumbersome dosing schedule, and its use is associated with the risk of thrombosis. Here, we present the computational design, engineering, and evaluation of FXa-variants that exhibit anticoagulation reversal activity in the presence of FXaDOACs. Our designs demonstrate low DOAC binding affinity, retain FXa-enzymatic activity and reduce the DOAC-associated bleeding by restoring hemostasis in mice treated with apixaban. Importantly, the FXaDOACs reversal agents we designed, unlike andexanet alfa, do not inhibit TFPI, and consequently, may have a safer thrombogenic profile.

The recent push for fair, trustworthy, and responsible Artificial Intelligence (AI) and Machine Learning (ML) systems have pushed for more explainable systems that are capable of explaining their predictions/decisions and inner workings. This led to the field of Explainable AI (XAI) going through an exponential growth in the past few years. XAI has been crucial in making AI/ML systems more comprehensible. However, XAI is limited to the model that it is being applied to, for both post-hoc or transparent models. Even though XAI can explain the decisions being made by the ML systems, these decisions are based on correlation and not causation. For applications such as tumor classification in the medical field, this can have serious consequences as people’s lives are affected. A potential solution for this challenge is the application of causal learning, which goes beyond the limitations of correlation for ML systems. Causal learning can generate analysis based on cause-and-effect relations within the data. This study compares the results of explanations given by post-hoc XAI systems to the causal features derived from causal graphs via causal discover for image datasets. We investigate how well XAI explanations/interpretations are able to identify the pertinent features within images. Causal graphs are generated for image datasets to extract the causal features that have a direct cause- and-effect relation with the label. These features are then compared to the features highlighted by XAI via feature relevance. The addition of causal learning for image datasets can aide in achieving fairness, bias detection, and mitigation to provide a robust and trustworthy system. We highlight the limitation of XAI tools such as LIME to make predictions based on physical features from images, whereas causal discovery can go beyond the simple pixel based perturbations to identify causal relations from image attributes.