Biblio

Learning-based systems have been shown to be vulnerable to evasion through adversarial data manipulation. These attacks have been studied under assumptions that the adversary has certain knowledge of either the target model internals, its training dataset or at least classification scores it assigns to input samples. In this paper, we investigate a much more constrained and realistic attack scenario wherein the target classifier is minimally exposed to the adversary, revealing only its final classification decision (e.g., reject or accept an input sample). Moreover, the adversary can only manipulate malicious samples using a blackbox morpher. That is, the adversary has to evade the targeted classifier by morphing malicious samples "in the dark". We present a scoring mechanism that can assign a real-value score which reflects evasion progress to each sample based on the limited information available. Leveraging on such scoring mechanism, we propose an evasion method – EvadeHC? and evaluate it against two PDF malware detectors, namely PDFRate and Hidost. The experimental evaluation demonstrates that the proposed evasion attacks are effective, attaining 100% evasion rate on the evaluation dataset. Interestingly, EvadeHC outperforms the known classifier evasion techniques that operate based on classification scores output by the classifiers. Although our evaluations are conducted on PDF malware classifiers, the proposed approaches are domain agnostic and are of wider application to other learning-based systems.

We study a sensor network setting in which samples are encrypted individually using different keys and maintained on a cloud storage. For large systems, e.g. those that generate several millions of samples per day, fine-grained sharing of encrypted samples is challenging. Existing solutions, such as Attribute-Based Encryption (ABE) and Key Aggregation Cryptosystem (KAC), can be utilized to address the challenge, but only to a certain extent. They are often computationally expensive and thus unlikely to operate at scale. We propose an algorithmic enhancement and two heuristics to improve KAC's key reconstruction cost, while preserving its provable security. The improvement is particularly significant for range and down-sampling queries – accelerating the reconstruction cost from quadratic to linear running time. Experimental study shows that for queries of size 32k samples, the proposed fast reconstruction techniques speed-up the original KAC by at least 90 times on range and down-sampling queries, and by eight times on general (arbitrary) queries. It also shows that at the expense of splitting the query into 16 sub-queries and correspondingly issuing that number of different aggregated keys, reconstruction time can be reduced by 19 times. As such, the proposed techniques make KAC more applicable in practical scenarios such as sensor networks or the Internet of Things.

We study a sensor network setting in which samples are encrypted individually using different keys and maintained on a cloud storage. For large systems, e.g. those that generate several millions of samples per day, fine-grained sharing of encrypted samples is challenging. Existing solutions, such as Attribute-Based Encryption (ABE) and Key Aggregation Cryptosystem (KAC), can be utilized to address the challenge, but only to a certain extent. They are often computationally expensive and thus unlikely to operate at scale. We propose an algorithmic enhancement and two heuristics to improve KAC's key reconstruction cost, while preserving its provable security. The improvement is particularly significant for range and down-sampling queries – accelerating the reconstruction cost from quadratic to linear running time. Experimental study shows that for queries of size 32k samples, the proposed fast reconstruction techniques speed-up the original KAC by at least 90 times on range and down-sampling queries, and by eight times on general (arbitrary) queries. It also shows that at the expense of splitting the query into 16 sub-queries and correspondingly issuing that number of different aggregated keys, reconstruction time can be reduced by 19 times. As such, the proposed techniques make KAC more applicable in practical scenarios such as sensor networks or the Internet of Things.

We study a sensor network setting in which samples are encrypted individually using different keys and maintained on a cloud storage. For large systems, e.g. those that generate several millions of samples per day, fine-grained sharing of encrypted samples is challenging. Existing solutions, such as Attribute-Based Encryption (ABE) and Key Aggregation Cryptosystem (KAC), can be utilized to address the challenge, but only to a certain extent. They are often computationally expensive and thus unlikely to operate at scale. We propose an algorithmic enhancement and two heuristics to improve KAC's key reconstruction cost, while preserving its provable security. The improvement is particularly significant for range and down-sampling queries – accelerating the reconstruction cost from quadratic to linear running time. Experimental study shows that for queries of size 32k samples, the proposed fast reconstruction techniques speed-up the original KAC by at least 90 times on range and down-sampling queries, and by eight times on general (arbitrary) queries. It also shows that at the expense of splitting the query into 16 sub-queries and correspondingly issuing that number of different aggregated keys, reconstruction time can be reduced by 19 times. As such, the proposed techniques make KAC more applicable in practical scenarios such as sensor networks or the Internet of Things.