With the increasing use of neural networks in critical systems, runtime monitoring becomes essential to reject unsafe predictions during inference. Various techniques have emerged to establish rejection scores that maximize the separability between the distributions of safe and unsafe predictions. The efficacy of these approaches is mostly evaluated using threshold-agnostic metrics, such as the area under the receiver operating characteristic curve. However, in real-world applications, an effective monitor also requires identifying a good threshold to transform these scores into meaningful binary decisions. Despite the pivotal importance of threshold optimization, this problem has received little attention. A few studies touch upon this question, but they typically assume that the runtime data distribution mirrors the training distribution, which is a strong assumption as monitors are supposed to safeguard a system against potentially unforeseen threats. In this work, we present rigorous experiments on various image datasets to investigate: 1. The effectiveness of monitors in handling unforeseen threats, which are not available during threshold adjustments. 2. Whether integrating generic threats into the threshold optimization scheme can enhance the robustness of monitors.

Whether we realize it or not, our digital lives and what we see on the internet are controlled and determined by algorithms and analytics. Through them, businesses learn what our preferences are and what we're drawn to in order to target us with information. The idea is to present us with information that is most relevant to us. In the same way, cybersecurity professionals are constantly faced with an enormous amount of threat data to sift through and prioritize on a daily basis. In fact, "too much data to analyze" is the number one obstacle inhibiting companies from defending against cyber threats according to the 2019 Cyberthreat Defense Report by CyberEdge.

Smart city improved the quality of life for the citizens by implementing information communication technology (ICT) such as the internet of things (IoT). Nevertheless, the smart city is a critical environment that needs to secure it is network and data from intrusions and attacks. This work proposes a hybrid deep learning (DL) model for cyber threat intelligence (CTI) to improve threats classification performance based on convolutional neural network (CNN) and quasi-recurrent neural network (QRNN). We use QRNN to provide a real-time threat classification model. The evaluation results of the proposed model compared to the state-of-the-art models show that the proposed model outperformed the other models. Therefore, it will help in classifying the smart city threats in a reasonable time.

Artificial intelligence in cybersecurity has recently made several headlines: "The Future of Cybersecurity: Artificial Intelligence" (Cybersecurity.CIOReview), "How AI is the Future of Cybersecurity" (Infosecurity Magazine) and "AI in Cyber Security Market to Grow at an Exorbitant Pace" (P&S Market Research). These headlines make seasoned cybersecurity professionals wary. We've seen other emerging technologies receive similar attention, and we've seen many of them fail to live up to their expectations. In this article, we will build a real-world perspective on AI in cybersecurity. We will explore where scepticism regarding AI in cybersecurity is justified, how the technology can provide tangible value, and what to look for in an AI-driven cybersecurity provider.

Threat data is no exception: Cybercriminals add to its abundance as they continuously up their game by tweaking old and creating new threats to evade detection. To address the vast amounts of threat data, security providers turn to machine learning to automate processes and improve security solutions. With the great diversity and volume of threat data available, machine learning is necessary to efficiently go through a dataset, learn from it, and help reinforce defenses against cyberthreats. The importance of the quantity of threat data is evident. But is data quantity the end all and be all of effective machine learning?

The threat landscape has moved far beyond programmers trying to show off their exploitative coding skills to their peers. Modern cybercriminals choose efficacy over spectacle and employ a variety of attack methodologies to breach network security. They leverage the most cutting-edge tech to launch swifter, more powerful, and highly sophisticated attacks. With advanced technologies such as machine learning and artificial intelligence now being integrated into cyber attack methodologies, security experts believe that 2018 could be the year that witnesses the first wave of attacks with true AI capabilities. This spells trouble for global businesses already struggling to deal with high attack volumes and multidimensional attack vectors.

Cybersecurity companies estimate that new malware variants are introduced at a daily rate of up to 390,000. With each hour that passes, at least 13,000 new files emerge. If you find these numbers staggering, that's because they are. Humans simply cannot keep up with them, which is why cybersecurity analysts are turning to artificial intelligence (AI) for help. Fighting the constantly evolving and morphing threat landscape requires a combination of detection and a single view of threat data, in addition to the traditional methods of signature-based malware detection and blocking.

I'm often asked what I like best about my job. One of my top answers is public speaking, learning and networking at security and technology events around the world. Besides giving press interviews or speeches on cyberthreats, I really enjoy moderating panels and leading executive roundtables with public- and private-sector leaders at security and technology events. I often get asked to be a moderator for a few sessions at SecureWorld Expo events, InfraGard Conferences and regional technology forums, such as the upcoming MidWest Technology Leaders event. During these panel sessions, the participants typically talk about a range of (hopefully intriguing) topics that include top cybercrime trends, cyberthreat intelligence, attracting and retaining cybertalent, big industry security breaches, internal security incidents or the always interesting (but overused question) "what's keeping you up at night?" Inevitably, security and technology topics include well known themes that I have written about such as ransomware, IoT botnets, cloud computing, smart cities, smartphone security, government CISO plans, securing the smart grid, end-user training, etc. Hopefully, we get beyond the problems and spend a few minutes on solutions.