A number of online services nowadays rely upon machine learning to extract valuable information from data collected in the wild. This exposes learning algorithms to the threat of data poisoning, i.e., a coordinate attack in which a fraction of the training data is controlled by the attacker and manipulated to subvert the learning process. To date, these attacks have been devised only against a limited class of binary learning algorithms, due to the inherent complexity of the gradient-based procedure used to optimize the poisoning points (a.k.a. adversarial training examples).
In this work, we first extend the definition of poisoning attacks to multi-class problems. We then propose a novel poisoning algorithm based on the idea of back-gradient optimization, i.e., to compute the gradient of interest through automatic differentiation, while also reversing the learning procedure to drastically reduce the attack complexity. Compared to current poisoning strategies, our approach is able to target a wider class of learning algorithms, trained with gradient-based procedures, including neural networks and deep learning architectures. We empirically evaluate its effectiveness on several application examples, including spam filtering, malware detection, and handwritten digit recognition. We finally show that, similarly to adversarial test examples, adversarial training examples can also be transferred across different learning algorithms.
Luis Muñoz-González, Battista Biggio, Ambra Demontis, Andrea Paudice, Vasin Wongrassamee, Emil C. Lupu, Fabio Roli. “Towards Poisoning Deep Learning Algorithms with Back-gradient Optimization.” Workshop on Artificial Intelligence and Security (AISec), 2017.
This work has been done in collaboration with the PRA Lab in the University of Cagliari, Italy.
Attack graphs provide compact representations of the attack paths an attacker can follow to compromise network resources from the analysis of network vulnerabilities and topology. These representations are a powerful tool for security risk assessment. Bayesian inference on attack graphs enables the estimation of the risk of compromise to the system’s
components given their vulnerabilities and interconnections and accounts for multi-step attacks spreading through the system. While static analysis considers the risk posture at rest, dynamic analysis also accounts for evidence of compromise, for example, from Security Information and Event Management software or forensic investigation. However, in this context, exact Bayesian inference techniques do not scale well. In this article, we show how Loopy Belief Propagation—an approximate inference technique—can be applied to attack graphs and that it scales linearly in the number of nodes for both static and dynamic analysis, making such analyses viable for larger networks. We experiment with different topologies and network clustering on synthetic Bayesian attack graphs with thousands of nodes to show that the algorithm’s accuracy is acceptable and that it converges to a stable solution. We compare sequential and parallel versions of Loopy Belief Propagation with exact inference techniques for both static and dynamic analysis, showing the advantages and gains of approximate inference techniques when scaling to larger attack graphs.
Luis Muñoz-González, Daniele Sgandurra, Andrea Paudice, Emil C. Lupu. “Efficient Attack Graph Analysis through Approximate Inference.” ACM Transactions on Privacy and Security, vol. 20(3), pp. 1-30, 2017.
Attack graphs are a powerful tool for security risk assessment by analysing network vulnerabilities and the paths attackers can use to compromise network resources. The uncertainty about the attacker’s behaviour makes Bayesian networks suitable to model attack graphs to perform static and dynamic analysis. Previous approaches have focused on the formalization of attack graphs into a Bayesian model rather than proposing mechanisms for their analysis. In this paper we propose to use efficient algorithms to make exact inference in Bayesian attack graphs, enabling the static and dynamic network risk assessments. To support the validity of our approach we have performed an extensive experimental evaluation on synthetic Bayesian attack graphs with different topologies, showing the computational advantages in terms of time and memory use of the proposed techniques when compared to existing approaches.
Luis Muñoz-González, Daniele Sgandurra, Martín Barrere, and Emil C. Lupu. “Exact Inference Techniques for the Analysis of Bayesian Attack Graphs.” IEEE Transactions on Dependable and Secure Computing (in press), 2017.
Attestation and measurements inspection are different but complementary approaches towards the same goal: ascertaining the integrity of sensor nodes in wireless sensor networks. In this paper we compare the benefits and drawbacks of both techniques and seek to determine how to best combine them. However, our study shows that no single solution exists, as each choice introduces changes in the measurements collection process, affects the attestation protocol, and gives a different balance between the high detection rate of attestation and the low power overhead of measurements inspection. Therefore, we propose three strategies that combine measurements inspection and attestation in different ways, and a way to choose between them based on the requirements of different applications. We analyse their performance both analytically and in a simulator. The results show that the combined strategies can achieve a detection rate close to attestation, in the range 96–99%, whilst keeping a power overhead close to measurements inspection, in the range 1–10%.
Vittorio P. Illiano, Rodrigo V. Steiner and Emil C. Lupu: Unity is strength!: combining attestation and measurements inspection to handle malicious data injections in WSNs.
ACM WiSec ’17 link (open access)
The paper “Enabling Data Sharing in Contextual Environments: Policy Representation and Analysis” was accepted at SACMAT 2017.
ACM Symposium on Access Control Models and Technologies (SACMAT 2017)
Authors: Erisa Karafili and Emil Lupu
Abstract: Internet of Things environments enable us to capture more and more data about the physical environment we live in and about ourselves. The data enable us to optimise resources, personalise services and offer unprecedented insights into our lives. However, to achieve these insights data need to be shared (and sometimes sold) between organisations imposing rights and obligations upon the sharing parties and in accordance with multiple layers of sometimes conflicting legislation at international, national and organisational levels. In this work, we show how such rules can be captured in a formal representation called “Data Sharing Agreements”. We introduce the use of abductive reasoning and argumentation based techniques to detect inconsistencies in the rules applicable and resolve them by assigning priorities to the rules. We show how through the use of argumentation based techniques use-cases taken from real life application are handled flexibly addressing trade-offs between confidentiality, privacy, availability and safety.
Wireless Sensor Networks carry a high risk of being compromised, as their deployments are often unattended, physically accessible and the wireless medium is difficult to secure. Malicious data injections take place when the sensed measurements are maliciously altered to trigger wrong and potentially dangerous responses. When many sensors are compromised, they can collude with each other to alter the measurements making such changes difficult to detect. Distinguishing between genuine and malicious measurements is even more difficult when significant variations may be introduced because of events, especially if more events occur simultaneously. We propose a novel methodology based on wavelet transform to detect malicious data injections, to characterise the responsible sensors, and to distinguish malicious interference from faulty behaviours. The results, both with simulated and real measurements, show that our approach is able to counteract sophisticated attacks, achieving a significant improvement over state-of-the-art approaches.
Vittorio P. Illiano, Luis Muñoz-González and Emil C. Lupu: Don’ t fool me!: Detection, Characterisation and Diagnosis of Spoofed and Masked Events in Wireless Sensor Networks.
To appear in IEEE Transactions on Dependable and Secure Computing
IEEE TNSM link (open access)
Attestation is a mechanism used by a trusted entity to validate the software integrity of an untrusted platform. Over the past few years, several attestation techniques have been proposed. While they all use variants of a challenge-response protocol, they make different assumptions about what an attacker can and cannot do. …
Recent statistics show that in 2015 more than 140 millions new malware samples have been found. Among these, a large portion is due to ransomware, the class of malware whose specific goal is to render the victim’s system unusable, in particular by encrypting important files, and then ask the user to pay a ransom to revert the damage. Several ransomware include sophisticated packing techniques, and are hence difficult to statically analyse. We present EldeRan, a machine learning approach for dynamically analysing and classifying ransomware. EldeRan monitors a set of actions performed by applications in their first phases of installation checking for characteristics signs of ransomware. Our tests over a dataset of 582 ransomware belonging to 11 families, and with 942 goodware applications, show that EldeRan achieves an area under the ROC curve of 0.995. Furthermore, EldeRan works without requiring that an entire ransomware family is available beforehand. These results suggest that dynamic analysis can support ransomware detection, since ransomware samples exhibit a set of characteristic features at run-time that are common across families, and that helps the early detection of new variants. We also outline some limitations of dynamic analysis for ransomware and propose possible solutions.
Daniele Sgandurra, Luis Muñoz-González, Rabih Mohsen, Emil C. Lupu. In ArXiv e-prints, arXiv:1609.03020, September 2016.
30th Annual IFIP WG 11.3 Working Conference on Data and Applications Security and Privacy (DBSec 2016)!
Authors: Daniele Sgandurra, Erisa Karafili and Emil Lupu.
Abstract: We propose a framework, called FATHoM (FormAlizing THreat Models), to define threat models for virtualized systems. For each component of a virtualized system, we specify a set of security proper- ties that defines its control responsibility, its vulnerability and protection states. Relations are used to represent how assumptions made about a component’s security state restrict the assumptions that can be made on the other components. FATHoM includes a set of rules to compute the derived security states from the assumptions and the components’ relations. A further set of relations and rules is used to define how to protect the derived vulnerable components. The resulting system is then analysed, among others, for consistency of the threat model. We have developed a tool that implements FATHoM, and have validated it with use-cases adapted from the literature.
Paper:Threat Model paper @ DBSec