Biblio

Machine Learning as a Service (MLaaS) is a growing paradigm in the Machine Learning (ML) landscape. More and more ML models are being uploaded to the cloud and made accessible from all over the world. Creating good ML models, however, can be expensive and the used data is often sensitive. Recently, Secure Multi-Party Computation (SMPC) protocols for MLaaS have been proposed, which protect sensitive user data and ML models at the expense of substantially higher computation and communication than plaintext evaluation. In this paper, we show that for a subset of ML models used in MLaaS, namely Support Vector Machines (SVMs) and Support Vector Regression Machines (SVRs) which have found many applications to classifying multimedia data such as texts and images, it is possible for adversaries to passively extract the private models even if they are protected by SMPC, using known and newly devised model extraction attacks. We show that our attacks are not only theoretically possible but also practically feasible and cheap, which makes them lucrative to financially motivated attackers such as competitors or customers. We perform model extraction attacks on the homomorphic encryption-based protocol for privacy-preserving SVR-based indoor localization by Zhang et al. (International Workshop on Security 2016). We show that it is possible to extract a highly accurate model using only 854 queries with the estimated cost of \$0.09 on the Amazon ML platform, and our attack would take only 7 minutes over the Internet. Also, we perform our model extraction attacks on SVM and SVR models trained on publicly available state-of-the-art ML datasets.

Plenty of methods applied in system identification, while those based on data-driven are increasingly popular. Usually we ignore the absence of outliers among the system to be modeled, but it is unreachable in reality. To improve the precision of identification towards system with outliers, advantageous approaches with robustness are needed. This study analyzes the superiority of weighted Least Square Support Vector Machine Regression (LS-SVMR) in the field of system identification under random outliers, and compare it with LS-SVMR mainly.

Recent indoor positioning systems that utilize channel state information (CSI) consider ideal scenarios to achieve high-accuracy performance in fingerprint matching. However, one essential component in achieving high accuracy is the collection of high-quality fingerprints. The quality of fingerprints may vary due to uncontrollable factors such as environment noise, interference, and hardware instability. In our paper, we propose a method for collecting high-quality fingerprints for indoor positioning. First, we have developed a logistic regression classifier based on gradient descent to evaluate the quality of the collected channel frequency response (CFR) samples. We employ the classifier to sift out poor CFR samples and only retain good ones as input to the positioning system. We discover that our classifier can achieve high classification accuracy from over thousands of CFR samples. We then evaluate the positioning accuracy based on two techniques: Time-Reversal Resonating Strength (TRRS) and Support Vector Machines (SVM). We find that the sifted fingerprints always result in better positioning performance. For example, an average percentage improvement of 114% for TRRS and 22% for SVM compared to that of unsifted fingerprints of the same 40-MHz effective bandwidth.

NeuronUnityIntgration2.0 (demo video is avilable at http://tiny.cc/u1lz6y) is a plugin for Unity which provides gesture recognition functionalities through the Perception Neuron motion capture suit. The system offers a recording mode, which guides the user through the collection of a dataset of gestures, and a recognition mode, capable of detecting the recorded actions in real time. Gestures are recognized by training Support Vector Machines directly within our plugin. We demonstrate the effectiveness of our application through an experimental evaluation on a newly collected dataset. Furthermore, external applications can exploit NeuronUnityIntgration2.0's recognition capabilities thanks to a set of exposed API.

We present a new method, called hyperplane folding, that increases the margin in Support Vector Machines (SVMs). Based on the location of the support vectors, the method splits the dataset into two parts, rotates one part of the dataset and then merges the two parts again. This procedure increases the margin as long as the margin is smaller than half of the shortest distance between any pair of data points from the two different classes. We provide an algorithm for the general case with n-dimensional data points. A small experiment with three folding iterations on 3-dimensional data points with non-linear relations shows that the margin does indeed increase and that the accuracy improves with a larger margin. The method can use any standard SVM implementation plus some basic manipulation of the data points, i.e., splitting, rotating and merging. Hyperplane folding also increases the interpretability of the data.

With the latest development of hydroelectric power generation system, the industrial control network system of hydroelectric power generation has undergone the transformation from the dedicated network, using proprietary protocols to an increasingly open network, adopting standard protocols, and increasing integration with hydroelectric power generation system. It generally believed that with the improvement of the smart grid, the future hydroelectric power generation system will rely more on the powerful network system. The general application of standardized communication protocol and intelligent electronic equipment in industrial control network provides a technical guarantee for realizing the intellectualization of hydroelectric power generation system but also brings about the network security problems that cannot be ignored. In order to solve the vulnerability of the system, we analyze and quantitatively evaluate the industrial control network of hydropower generation as a whole, and propose a set of attack and defense strategies. The method of vulnerability assessment with high diversity score proposed by us avoids the indifference of different vulnerability score to the greatest extent. At the same time, we propose an optimal attack and defense decision algorithm, which generates the optimal attack and defense strategy. The work of this paper can distinguish the actual hazards of vulnerable points more effectively.

Stealth malware, a representative tool of advanced persistent threat (APT) attacks, in particular poses an increased threat to cyber-physical systems (CPS). Due to the use of stealthy and evasive techniques (e.g., zero-day exploits, obfuscation techniques), stealth malwares usually render conventional heavyweight countermeasures (e.g., exploits patching, specialized ant-malware program) inapplicable. Light-weight countermeasures (e.g., containment techniques), on the other hand, can help retard the spread of stealth malwares, but the ensuing side effects might violate the primary safety requirement of CPS. Hence, defenders need to find a balance between the gain and loss of deploying light-weight countermeasures. To address this challenge, we model the persistent anti-malware process as a shortest-path tree interdiction (SPTI) Stackelberg game, and safety requirements of CPS are introduced as constraints in the defender's decision model. Specifically, we first propose a static game (SSPTI), and then extend it to a multi-stage dynamic game (DSPTI) to meet the need of real-time decision making. Both games are modelled as bi-level integer programs, and proved to be NP-hard. We then develop a Benders decomposition algorithm to achieve the Stackelberg Equilibrium of SSPTI. Finally, we design a model predictive control strategy to solve DSPTI approximately by sequentially solving an approximation of SSPTI. The extensive simulation results demonstrate that the proposed dynamic defense strategy can achieve a balance between fail-secure ability and fail-safe ability while retarding the stealth malware propagation in CPS.

By accurately measuring risk for enterprise networks, attack graphs allow network defenders to understand the most critical threats and select the most effective countermeasures. This paper describes substantial enhancements to the NetSPA attack graph system required to model additional present-day threats (zero-day exploits and client-side attacks) and countermeasures (intrusion prevention systems, proxy firewalls, personal firewalls, and host-based vulnerability scans). Point-to-point reachability algorithms and structures were extensively redesigned to support "reverse" reachability computations and personal firewalls. Host-based vulnerability scans are imported and analyzed. Analysis of an operational network with 84 hosts demonstrates that client-side attacks pose a serious threat. Experiments on larger simulated networks demonstrated that NetSPA's previous excellent scaling is maintained. Less than two minutes are required to completely analyze a four-enclave simulated network with more than 40,000 hosts protected by personal firewalls.

Zero-day attacks continue to challenge the enterprise network security defense. A zero-day attack path is formed when a multi-step attack contains one or more zero-day exploits. Detecting zero-day attack paths in time could enable early disclosure of zero-day threats. In this paper, we propose a probabilistic approach to identify zero-day attack paths and implement a prototype system named ZePro. An object instance graph is first built from system calls to capture the intrusion propagation. To further reveal the zero-day attack paths hiding in the instance graph, our system constructs an instance-graph-based Bayesian network. By leveraging intrusion evidence, the Bayesian network can quantitatively compute the probabilities of object instances being infected. The object instances with high infection probabilities reveal themselves and form the zero-day attack paths. The experiment results show that our system can effectively identify zero-day attack paths.

Computer networks are overwhelmed by self propagating malware (worms, viruses, trojans). Although the number of security vulnerabilities grows every day, not the same thing can be said about the number of defense methods. But the most delicate problem in the information security domain remains detecting unknown attacks known as zero-day attacks. This paper presents methods for isolating the malicious traffic by using a honeypot system and analyzing it in order to automatically generate attack signatures for the Snort intrusion detection/prevention system. The honeypot is deployed as a virtual machine and its job is to log as much information as it can about the attacks. Then, using a protected machine, the logs are collected remotely, through a safe connection, for analysis. The challenge is to mitigate the risk we are exposed to and at the same time search for unknown attacks.

A3 is an execution management environment that aims to make network-facing applications and services resilient against zero-day attacks. A3 recently underwent two adversarial evaluations of its defensive capabilities. In one, A3 defended an App Store used in a Capture the Flag (CTF) tournament, and in the other, a tactically relevant network service in a red team exercise. This paper describes the A3 defensive technologies evaluated, the evaluation results, and the broader lessons learned about evaluations for technologies that seek to protect critical systems from zero-day attacks.

With technology at our fingertips waiting to be exploited, the past decade saw the revolutionizing Human Computer Interactions. The ease with which a user could interact was the Unique Selling Proposition (USP) of a sales team. Human Computer Interactions have many underlying parameters like Data Visualization and Presentation as some to deal with. With the race, on for better and faster presentations, evolved many frameworks to be widely used by all software developers. As the need grew for user friendly applications, more and more software professionals were lured into the front-end sophistication domain. Application frameworks have evolved to such an extent that with just a few clicks and feeding values as per requirements we are able to produce a commercially usable application in a few minutes. These frameworks generate quantum lines of codes in minutes which leaves a contrail of bugs to be discovered in the future. We have also succumbed to the benchmarking in Software Quality Metrics and have made ourselves comfortable with buggy software's to be rectified in future. The exponential evolution in the cyber domain has also attracted attackers equally. Average human awareness and knowledge has also improved in the cyber domain due to the prolonged exposure to technology for over three decades. As the attack sophistication grows and zero day attacks become more popular than ever, the suffering end users only receive remedial measures in spite of the latest Antivirus, Intrusion Detection and Protection Systems installed. We designed a software to display the complete services and applications running in users Operating System in the easiest perceivable manner aided by Computer Graphics and Data Visualization techniques. We further designed a study by empowering the fence sitter users with tools to actively participate in protecting themselves from threats. The designed threats had impressions from the complete threat canvas in some form or other restricted to systems functioning. Network threats and any sort of packet transfer to and from the system in form of threat was kept out of the scope of this experiment. We discovered that end users had a good idea of their working environment which can be used exponentially enhances machine learning for zero day threats and segment the unmarked the vast threat landscape faster for a more reliable output.

We present a basic tool for zero day attack signature extraction. Given two large sets of messages, \$P\$ the messages captured in the network at peacetime i.e., mostly legitimate traffic and \$A\$ the...

Advanced Persistent Threats (APT) are highly targeted and sophisticated multi-stage attacks, utilizing zero day or near zero-day malware. Directed at internetworked computer users in the workplace, their growth and prevalence can be attributed to both socio (human) and technical (system weaknesses and inadequate cyber defenses) vulnerabilities. While many APT attacks incorporate a blend of socio-technical vulnerabilities, academic research and reported incidents largely depict the user as the prominent contributing factor that can weaken the layers of technical security in an organization. In this paper, our objective is to explore multiple dimensions of socio factors (non-technical vulnerabilities) that contribute to the success of APT attacks in organizations. Expert interviews were conducted with senior managers, working in government and private organizations in the United Arab Emirates (UAE) over a period of four years (2014 to 2017). Contrary to common belief that socio factors derive predominately from user behavior, our study revealed two new dimensions of socio vulnerabilities, namely the role of organizational management, and environmental factors which also contribute to the success of APT attacks. We show that the three dimensions postulated in this study can assist Managers and IT personnel in organizations to implement an appropriate mix of socio-technical countermeasures for APT threats.

In order to examine malicious activity that occurs in a network or a system, intrusion detection system is used. Intrusion Detection is software or a device that scans a system or a network for a distrustful activity. Due to the growing connectivity between computers, intrusion detection becomes vital to perform network security. Various machine learning techniques and statistical methodologies have been used to build different types of Intrusion Detection Systems to protect the networks. Performance of an Intrusion Detection is mainly depends on accuracy. Accuracy for Intrusion detection must be enhanced to reduce false alarms and to increase the detection rate. In order to improve the performance, different techniques have been used in recent works. Analyzing huge network traffic data is the main work of intrusion detection system. A well-organized classification methodology is required to overcome this issue. This issue is taken in proposed approach. Machine learning techniques like Support Vector Machine (SVM) and Naïve Bayes are applied. These techniques are well-known to solve the classification problems. For evaluation of intrusion detection system, NSL- KDD knowledge discovery Dataset is taken. The outcomes show that SVM works better than Naïve Bayes. To perform comparative analysis, effective classification methods like Support Vector Machine and Naive Bayes are taken, their accuracy and misclassification rate get calculated.

Generating public randomness has been significantly demanding and also challenging, especially after the introduction of the Blockchain Technology. Lotteries, smart contracts, and random audits are examples where the reliability of the randomness source is a vital factor. We demonstrate a system of random number generation service for generating fair, tamper-resistant, and verifiable random numbers. Our protocol together with this system is an R&D project aiming at providing a decentralized solution to random number generation by leveraging the blockchain technology along with long-lasting cryptographic primitives including homomorphic encryption, verifiable random functions. The system decentralizes the process of generating random numbers by combining each party's favored value to obtain the final random numbers. Our novel idea is to force each party to encrypt his contribution before making it public. With the help of homomorphic encryption, all encrypted contribution can be combined without performing any decryption. The solution has achieved the properties of unpredictability, tamper-resistance, and public-verifiability. In addition, it only offers a linear overall complexity with respect to the number of parties on the network, which permits great scalability.

Distributed applications cannot assume that their security policies will be enforced on untrusted hosts. Trusted execution environments (TEEs) combined with cryptographic mechanisms enable execution of known code on an untrusted host and the exchange of confidential and authenticated messages with it. TEEs do not, however, establish the trustworthiness of code executing in a TEE. Thus, developing secure applications using TEEs requires specialized expertise and careful auditing. This paper presents DFLATE, a core security calculus for distributed applications with TEEs. DFLATE offers high-level abstractions that reflect both the guarantees and limitations of the underlying security mechanisms they are based on. The accuracy of these abstractions is exhibited by asymmetry between confidentiality and integrity in our formal results: DFLATE enforces a strong form of noninterference for confidentiality, but only a weak form for integrity. This reflects the asymmetry of the security guarantees of a TEE: a malicious host cannot access secrets in the TEE or modify its contents, but they can suppress or manipulate the sequence of its inputs and outputs. Therefore DFLATE cannot protect against the suppression of high-integrity messages, but when these messages are delivered, their contents cannot have been influenced by an attacker.

Advancing network technologies allows the setup of two-way communication links between energy providers and consumers. These developing technologies aim to enhance grid reliability and energy efficiency in smart grids. To achieve this goal, energy usage reports from consumers are required to be both trustworthy and confidential. In this paper, we construct a new data aggregation scheme in smart grids based on a homomorphic encryption algorithm. In the constructed scheme, obedient consumers who follow the instruction can prove its ajustment using a range proof protocol. Additionally, we propose a new identity-based signature algorithm in order to ensure authentication and integrity of the constructed scheme. By using this signature algorithm, usage reports are verified in real time. Extensive simulations demonstrate that our scheme outperforms other data aggregation schemes.

As cloud services enter the Internet market, cloud security issues are gradually exposed. In the era of knowledge economy, the unique potential value of big data is being gradually explored. However, the control of data security is facing many challenges. According to the development status and characteristics of database within the cloud environment, this paper preliminary studies on the database security risks faced by the “three-clouds” of State Grid Corporation of China. Based on the mature standardization of information security, this paper deeply studies the database security requirements of cloud environment, and six-step method for cloud database protection is presented, which plays an important role in promoting development of security work for the cloud database. Four key technologies of cloud database security protection are introduced, including database firewall technology, sensitive data encryption, production data desensitization, and database security audit technology. It is helpful to the technology popularization of the grade protection in the security of the cloud database, and plays a great role in the construction of the security of the state grid.

We propose an approach for allowing data owners to trade their data in digital data market scenarios, while keeping control over them. Our solution is based on a combination of selective encryption and smart contracts deployed on a blockchain, and ensures that only authorized users who paid an agreed amount can access a data item. We propose a safe interaction protocol for regulating the interplay between a data owner and subjects wishing to purchase (a subset of) her data, and an audit process for counteracting possible misbehaviors by any of the interacting parties. Our solution aims to make a step towards the realization of data market platforms where owners can benefit from trading their data while maintaining control.

With the advent of the big data era, information systems have exhibited some new features, including boundary obfuscation, system virtualization, unstructured and diversification of data types, and low coupling among function and data. These features not only lead to a big difference between big data technology (DT) and information technology (IT), but also promote the upgrading and evolution of network security technology. In response to these changes, in this paper we compare the characteristics between IT era and DT era, and then propose four DT security principles: privacy, integrity, traceability, and controllability, as well as active and dynamic defense strategy based on "propagation prediction, audit prediction, dynamic management and control". We further discuss the security challenges faced by DT and the corresponding assurance strategies. On this basis, the big data security technologies can be divided into four levels: elimination, continuation, improvement, and innovation. These technologies are analyzed, combed and explained according to six categories: access control, identification and authentication, data encryption, data privacy, intrusion prevention, security audit and disaster recovery. The results will support the evolution of security technologies in the DT era, the construction of big data platforms, the designation of security assurance strategies, and security technology choices suitable for big data.

Many enterprises and individuals prefer to outsource data to public cloud via various pricing approaches. One of the most widely-used approaches is the pay-as-you-go model, where the data owner hires public cloud to share data with data consumers, and only pays for the actually consumed services. To realize controllable and secure data sharing, ciphertext-policy attribute-based encryption (CP-ABE) is a suitable solution, which can provide fine-grained access control and encryption functionalities simultaneously. But there are some serious challenges when applying CP-ABE in pay-as-you-go. Firstly, the decryption cost in ABE is too heavy for data consumers. Secondly, ABE ciphertexts probably suffer distributed denial of services (DDoS) attacks, but there is no solution that can eliminate the security risk. At last, the data owner should audit resource consumption to guarantee the transparency of charge, while the existing method is inefficient. In this work, we propose a general construction named fully accountable ABE (FA-ABE), which simultaneously solves all the challenges by supporting all-sided accountability in the pay-as-you-go model. We formally define the security model and prove the security in the standard model. Also, we implement an instantiate construction with the self-developed library libabe. The experiment results indicate the efficiency and practicality of our construction.

Security auditing allows cloud tenants to verify the compliance of cloud infrastructure with respect to desirable security properties, e.g., whether a tenant's virtual network is properly isolated from other tenants' networks. However, the input to such an auditing task, such as the detailed topology of the underlying cloud infrastructure, typically contains sensitive information which a cloud provider may be reluctant to hand over to a third party auditor. Additionally, auditing results intended for one tenant may inadvertently reveal private information about other tenants, e.g., another tenant's VM is reachable due to a misconfiguration. How to anonymize both the input data and the auditing results in order to prevent such information leakage is a novel challenge that has received little attention. Directly applying most of the existing anonymization techniques to such a context would either lead to insufficient protection or render the data unsuitable for auditing. In this paper, we propose SegGuard, a novel anonymization approach that prevents cross-tenant information leakage through per-tenant encryption, and prevents information leakage to auditors through hiding real input segments among fake ones; in addition, applying property-preserving encryption in an innovative way enables SegGuard to preserve the data utility for auditing while mitigating semantic attacks. We implement SegGuard based on OpenStack, and evaluate its effectiveness and overhead using both synthetic and real data. Our experimental results demonstrate that SegGuard can reduce the information leakage to a negligible level (e.g., less than 1% for an adversary with 50% pre-knowledge) with a practical response time (e.g., 62 seconds to anonymize a cloud infrastructure with 25,000 virtual machines).

Data security is a major requirement of smart meter communication to control server through Advanced Metering infrastructure. Easy access of smart meters and multi-faceted nature of AMI communication network are the main reasons of smart meter facing large number of attacks. The different topology, bandwidth and heterogeneity in communication network prevent the existing security mechanisms in satisfying the security requirements of smart meter. Hence, advanced security mechanisms are essential to encrypt smart meter data before transmitting to control server. The emerging biocryptography technique has several advantages over existing techniques and is most suitable for providing security to communication of low processing devices like smart meter. In this paper, a lightweight encryption scheme using DNA sequence with suitable key management scheme is proposed for secure communication of smart meter in an efficient way. The proposed 2-phase DNA cryptography provides confidentiality and integrity to transmitted data and the authentication of keys is attained by exchanging through Diffie Hellman scheme. The strength of proposed encryption scheme is analyzed and its efficiency is evaluated by simulating an AMI communication network using Simulink/Matlab. Comparison of simulation results with various techniques show that the proposed scheme is suitable for secure communication of smart meter data.

DNA cryptography becomes a burgeoning new area of study along with the fast-developing of DNA computing and modern cryptography. Point-doubling, point-addition and point-multiplication are three fundamental point-operations to construct encryption protocols in some cryptosystem over mathematical curves such as elliptic curves and conic curves. This paper proposes a DNA computing model to calculate point-doubling in conic curves cryptosystem over finite held GF(2n). By decomposing and rearranging the computing steps of point-doubling, the assembly process could be fulfilled by using 8 different types of computation tiles performing different functions with 1097 encoding ways. This model could also figure out point-multiplication if its coefficient is 2k. The assembly time complexity is 2kn+n-k-1, and the space complexity is k2n2+kn2-k2n.