Luigi Vincenzo Mancini

Sabrine Ennaji, Elhadj Benkhelifa, Luigi Vincenzo Mancini

Machine learning based intrusion detection systems are increasingly targeted by black box adversarial attacks, where attackers craft evasive inputs using indirect feedback such as binary outputs or behavioral signals like response time and resource usage. While several defenses have been proposed, including input transformation, adversarial training, and surrogate detection, they often fall short in practice. Most are tailored to specific attack types, require internal model access, or rely on static mechanisms that fail to generalize across evolving attack strategies. Furthermore, defenses such as input transformation can degrade intrusion detection system performance, making them unsuitable for real time deployment. To address these limitations, we propose Adaptive Feature Poisoning, a lightweight and proactive defense mechanism designed specifically for realistic black box scenarios. Adaptive Feature Poisoning assumes that probing can occur silently and continuously, and introduces dynamic and context aware perturbations to selected traffic features, corrupting the attacker feedback loop without impacting detection capabilities. The method leverages traffic profiling, change point detection, and adaptive scaling to selectively perturb features that an attacker is likely exploiting, based on observed deviations. We evaluate Adaptive Feature Poisoning against multiple realistic adversarial attack strategies, including silent probing, transferability based attacks, and decision boundary based attacks. The results demonstrate its ability to confuse attackers, degrade attack effectiveness, and preserve detection performance. By offering a generalizable, attack agnostic, and undetectable defense, Adaptive Feature Poisoning represents a significant step toward practical and robust adversarial resilience in machine learning based intrusion detection systems.

Ying Yuan, Cristiano Alex Rado, Giovanni Apruzzese et al.

Phishing webpages are continuously polluting the Web. Plenty of countermeasures have been proposed and the most advanced techniques leverage machine-learning methods that infer whether a webpage is benign or not by inspecting its visual representation. Yet, despite the demonstrated effectiveness of such detection methods, this class of defenses is, by design, susceptible to a kind of subtle-but-cheap timing-based attacks which -- worryingly, and perhaps surprisingly -- have never been investigated so far. Such an oversight questions the overall reliability of these defenses in the wild. First, we show that timing-based evasion attacks have not been accounted for by prior work on visual phishing websites detectors. Then, we elucidate the intrinsic vulnerability of these detectors: they can be bypassed by delaying the rendering of webpage elements. Practically, these detectors must compute the visual similarity between a target webpage and a known legitimate one. This requires taking a "snapshot" of the target webpage before the similarity computation. Attackers can deliberately delay the rendering of key elements, such as the logo, so that these elements appear fully only after the snapshot has been taken. This simple tactic misleads the visual-similarity module, leading the system to incorrectly classify the phishing page as benign. We empirically show that state-of-the-art detectors can be completely defeated (detection rate dropping from 100% to 0%) by employing easy-to-apply problem-space techniques such as curtain effects. We also carry out a user study, evaluating the effectiveness of these attacks against real humans, and find that end users are unable to reliably identify our "perturbations" (p<.05). Finally, we propose mitigations, including a browser-extension that, without making any call to remote services, warns users that they may have landed on a phishing webpage.

Nino Vincenzo Verde, Giuseppe Ateniese, Emanuele Gabrielli et al.

It is generally recognized that the traffic generated by an individual connected to a network acts as his biometric signature. Several tools exploit this fact to fingerprint and monitor users. Often, though, these tools assume to access the entire traffic, including IP addresses and payloads. This is not feasible on the grounds that both performance and privacy would be negatively affected. In reality, most ISPs convert user traffic into NetFlow records for a concise representation that does not include, for instance, any payloads. More importantly, large and distributed networks are usually NAT'd, thus a few IP addresses may be associated to thousands of users. We devised a new fingerprinting framework that overcomes these hurdles. Our system is able to analyze a huge amount of network traffic represented as NetFlows, with the intent to track people. It does so by accurately inferring when users are connected to the network and which IP addresses they are using, even though thousands of users are hidden behind NAT. Our prototype implementation was deployed and tested within an existing large metropolitan WiFi network serving about 200,000 users, with an average load of more than 1,000 users simultaneously connected behind 2 NAT'd IP addresses only. Our solution turned out to be very effective, with an accuracy greater than 90%. We also devised new tools and refined existing ones that may be applied to other contexts related to NetFlow analysis.