Visible to the public Biblio

Found 333 results

Filters: Keyword is predictive metrics  [Clear All Filters]
2021-05-03
Das, Arnab, Briggs, Ian, Gopalakrishnan, Ganesh, Krishnamoorthy, Sriram, Panchekha, Pavel.  2020.  Scalable yet Rigorous Floating-Point Error Analysis. SC20: International Conference for High Performance Computing, Networking, Storage and Analysis. :1–14.
Automated techniques for rigorous floating-point round-off error analysis are a prerequisite to placing important activities in HPC such as precision allocation, verification, and code optimization on a formal footing. Yet existing techniques cannot provide tight bounds for expressions beyond a few dozen operators-barely enough for HPC. In this work, we offer an approach embedded in a new tool called SATIHE that scales error analysis by four orders of magnitude compared to today's best-of-class tools. We explain how three key ideas underlying SATIHE helps it attain such scale: path strength reduction, bound optimization, and abstraction. SATIHE provides tight bounds and rigorous guarantees on significantly larger expressions with well over a hundred thousand operators, covering important examples including FFT, matrix multiplication, and PDE stencils.
Naik, Nikhil, Nuzzo, Pierluigi.  2020.  Robustness Contracts for Scalable Verification of Neural Network-Enabled Cyber-Physical Systems. 2020 18th ACM-IEEE International Conference on Formal Methods and Models for System Design (MEMOCODE). :1–12.
The proliferation of artificial intelligence based systems in all walks of life raises concerns about their safety and robustness, especially for cyber-physical systems including multiple machine learning components. In this paper, we introduce robustness contracts as a framework for compositional specification and reasoning about the robustness of cyber-physical systems based on neural network (NN) components. Robustness contracts can encompass and generalize a variety of notions of robustness which were previously proposed in the literature. They can seamlessly apply to NN-based perception as well as deep reinforcement learning (RL)-enabled control applications. We present a sound and complete algorithm that can efficiently verify the satisfaction of a class of robustness contracts on NNs by leveraging notions from Lagrangian duality to identify system configurations that violate the contracts. We illustrate the effectiveness of our approach on the verification of NN-based perception systems and deep RL-based control systems.
Herber, Paula, Liebrenz, Timm.  2020.  Dependence Analysis and Automated Partitioning for Scalable Formal Analysis of SystemC Designs. 2020 18th ACM-IEEE International Conference on Formal Methods and Models for System Design (MEMOCODE). :1–6.
Embedded systems often consist of deeply intertwined hardware and software components. At the same time, they are often used in safety-critical applications, where an error may result in enormous costs or even loss of human lives. Existing verification techniques that show the absence of errors do not scale well for complex integrated HW/SW systems. In this paper, we present a dependence analysis and automated partitioning approach for the formal analysis of HW/SW codesigns that are modeled in SystemC. The key idea of our approach is threefold: first, we partition a given system into loosely coupled submodels. Second, we analyze the dependences between these submodels and compute an abstract verification interface for each of them, which captures all possible influences of all other submodels. Third, we verify global properties of the overall system by verifying them separately for each subsystem. We demonstrate that our approach significantly reduces verification times and increases scalability with results for an anti-lock braking system.
Sharma, Mohit, Strathman, Hunter J., Walker, Ross M..  2020.  Verification of a Rapidly Multiplexed Circuit for Scalable Action Potential Recording. 2020 IEEE International Symposium on Circuits and Systems (ISCAS). :1–1.
This report presents characterizations of in vivo neural recordings performed with a CMOS multichannel chip that uses rapid multiplexing directly at the electrodes, without any pre-amplification or buffering. Neural recordings were taken from a 16-channel microwire array implanted in rodent cortex, with comparison to a gold-standard commercial bench-top recording system. We were able to record well-isolated threshold crossings from 10 multiplexed electrodes and typical local field potential waveforms from 16, with strong agreement with the standard system (average SNR = 2.59 and 3.07 respectively). For 10 electrodes, the circuit achieves an effective area per channel of 0.0077 mm2, which is \textbackslashtextgreater5× smaller than typical multichannel chips. Extensive characterizations of noise and signal quality are presented and compared to fundamental theory, as well as results from in vivo and in vitro experiments. By demonstrating the validation of rapid multiplexing directly at the electrodes, this report confirms it as a promising approach for reducing circuit area in massively-multichannel neural recording systems, which is crucial for scaling recording site density and achieving large-scale sensing of brain activity with high spatiotemporal resolution.
Adelt, Peer, Koppelmann, Bastian, Mueller, Wolfgang, Scheytt, Christoph.  2020.  A Scalable Platform for QEMU Based Fault Effect Analysis for RISC-V Hardware Architectures. MBMV 2020 - Methods and Description Languages for Modelling and Verification of Circuits and Systems; GMM/ITG/GI-Workshop. :1–8.
Fault effect simulation is a well-established technique for the qualification of robust embedded software and hardware as required by different safety standards. Our article introduces a Virtual Prototype based approach for the fault analysis and fast simulation of a set of automatically generated and target compiled software programs. The approach scales to different RISC-V ISA standard subset configurations and is based on an instruction and hardware register coverage for automatic fault injections of permanent and transient bitflips. The analysis of each software binary evaluates its opcode type and register access coverage including the addressed memory space. Based on this information dedicated sets of fault injected hardware models, i.e., mutants, are generated. The simulation of all mutants conducted with the different binaries finally identifies the cases with a normal termination though executed on a faulty hardware model. They are identified as a subject for further investigations and improvements by the implementation of additional hardware or software safety countermeasures. Our final evaluation results with automatic C code generation, compilation, analysis, and simulation show that QEMU provides an adequate efficient platform, which also scales to more complex scenarios.
Shen, Shen, Tedrake, Russ.  2020.  Sampling Quotient-Ring Sum-of-Squares Programs for Scalable Verification of Nonlinear Systems. 2020 59th IEEE Conference on Decision and Control (CDC). :2535–2542.
This paper presents a novel method, combining new formulations and sampling, to improve the scalability of sum-of-squares (SOS) programming-based system verification. Region-of-attraction approximation problems are considered for polynomial, polynomial with generalized Lur'e uncertainty, and rational trigonometric multi-rigid-body systems. Our method starts by identifying that Lagrange multipliers, traditionally heavily used for S-procedures, are a major culprit of creating bloated SOS programs. In light of this, we exploit inherent system properties-continuity, convexity, and implicit algebraic structure-and reformulate the problems as quotient-ring SOS programs, thereby eliminating all the multipliers. These new programs are smaller, sparser, less constrained, yet less conservative. Their computation is further improved by leveraging a recent result on sampling algebraic varieties. Remarkably, solution correctness is guaranteed with just a finite (in practice, very small) number of samples. Altogether, the proposed method can verify systems well beyond the reach of existing SOS-based approaches (32 states); on smaller problems where a baseline is available, it computes tighter solution 2-3 orders of magnitude faster.
Wu, Shanglun, Yuan, Yujie, Kar, Pushpendu.  2020.  Lightweight Verification and Fine-grained Access Control in Named Data Networking Based on Schnorr Signature and Hash Functions. 2020 IEEE 20th International Conference on Communication Technology (ICCT). :1561–1566.
Named Data Networking (NDN) is a new kind of architecture for future Internet, which is exactly satisfied with the rapidly increasing mobile requirement and information-depended applications that dominate today's Internet. However, the current verification-data accessed system is not safe enough to prevent data leakage because no strongly method to resist any device or user to access it. We bring up a lightweight verification based on hash functions and a fine-grained access control based on Schnorr Signature to address the issue seamlessly. The proposed scheme is scalable and protect data confidentiality in a NDN network.
Paulsen, Brandon, Wang, Jingbo, Wang, Jiawei, Wang, Chao.  2020.  NEURODIFF: Scalable Differential Verification of Neural Networks using Fine-Grained Approximation. 2020 35th IEEE/ACM International Conference on Automated Software Engineering (ASE). :784–796.
As neural networks make their way into safety-critical systems, where misbehavior can lead to catastrophes, there is a growing interest in certifying the equivalence of two structurally similar neural networks - a problem known as differential verification. For example, compression techniques are often used in practice for deploying trained neural networks on computationally- and energy-constrained devices, which raises the question of how faithfully the compressed network mimics the original network. Unfortunately, existing methods either focus on verifying a single network or rely on loose approximations to prove the equivalence of two networks. Due to overly conservative approximation, differential verification lacks scalability in terms of both accuracy and computational cost. To overcome these problems, we propose NEURODIFF, a symbolic and fine-grained approximation technique that drastically increases the accuracy of differential verification on feed-forward ReLU networks while achieving many orders-of-magnitude speedup. NEURODIFF has two key contributions. The first one is new convex approximations that more accurately bound the difference of two networks under all possible inputs. The second one is judicious use of symbolic variables to represent neurons whose difference bounds have accumulated significant error. We find that these two techniques are complementary, i.e., when combined, the benefit is greater than the sum of their individual benefits. We have evaluated NEURODIFF on a variety of differential verification tasks. Our results show that NEURODIFF is up to 1000X faster and 5X more accurate than the state-of-the-art tool.
Zalasiński, Marcin, Cpałka, Krzysztof, Łapa, Krystian.  2020.  An interpretable fuzzy system in the on-line signature scalable verification. 2020 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE). :1–9.
This paper proposes new original solutions for the use of interpretable flexible fuzzy systems for identity verification based on an on-line signature. Such solutions must be scalable because the verification of the identity of each user must be carried out independently of one another. In addition, a large number of system users limit the possibilities of iterative system learning. An important issue is the ability to interpret the system rules because it explains how the similarity of test signatures to reference signature templates is assessed. In this paper, we propose an approach that meets all of the above requirements and works effectively for the on-line signatures' database used in the simulations.
Takita, Yutaka, Miyabe, Masatake, Tomonaga, Hiroshi, Oguchi, Naoki.  2020.  Scalable Impact Range Detection against Newly Added Rules for Smart Network Verification. 2020 IEEE 44th Annual Computers, Software, and Applications Conference (COMPSAC). :1471–1476.
Technological progress in cloud networking, 5G networks, and the IoT (Internet of Things) are remarkable. In addition, demands for flexible construction of SoEs (Systems on Engagement) for various type of businesses are increasing. In such environments, dynamic changes of network rules, such as access control (AC) or packet forwarding, are required to ensure function and security in networks. On the other hand, it is becoming increasingly difficult to grasp the exact situation in such networks by utilizing current well-known network verification technologies since a huge number of network rules are complexly intertwined. To mitigate these issues, we have proposed a scalable network verification approach utilizing the concept of "Packet Equivalence Class (PEC)," which enable precise network function verification by strictly recognizing the impact range of each network rule. However, this approach is still not scalable for very large-scale networks which consist of tens of thousands of routers. In this paper, we enhanced our impact range detection algorithm for practical large-scale networks. Through evaluation in the network with more than 80,000 AC rules, we confirmed that our enhanced algorithm can achieve precise impact range detection in under 600 seconds.
Le, Son N., Srinivasan, Sudarshan K., Smith, Scott C..  2020.  Exploiting Dual-Rail Register Invariants for Equivalence Verification of NCL Circuits. 2020 IEEE 63rd International Midwest Symposium on Circuits and Systems (MWSCAS). :21–24.
Equivalence checking is one of the most scalable and useful verification techniques in industry. NULL Convention Logic (NCL) circuits utilize dual-rail signals (i.e., two wires to represent one bit of DATA), where the wires are inverses of each other during a DATA wavefront. In this paper, a technique that exploits this invariant at NCL register boundaries is proposed to improve the efficiency of equivalence verification of NCL circuits.
Raj A.G.R., Rahul, Sunitha, R., Prasad, H.B..  2020.  Mitigating DDoS Flooding Attacks with Dynamic Path Identifiers in Wireless Network. 2020 Second International Conference on Inventive Research in Computing Applications (ICIRCA). :869–874.
The usage of wireless devices is increased from last decade due to its reliable, fast and easy transfer of data. Ensuring the security to these networks is a crucial thing. There are several types of network attacks, in this paper, DDoS attacks on networks and techniques, consequences, effects and prevention methods are focused on. The DDoS attack is carried out by multiple attackers on a system which floods the system with a greater number of incoming requests to the system. The destination system cannot immediately respond to the huge requests, due to this server crashes or halts. To detect, or to avoid such scenarios Intrusion prevention system is designed. The IPS block the network attacker at its first hop and thus reduce the malicious traffic near its source. Intrusion detection system prevents the attack without the prior knowledge of the attacker. The attack is detected at the router side and path is changed to transfer the files. The proposed model is designed to obtain the dynamic path for efficient transmission in wireless neworks.
Gelenbe, Erol.  2020.  Machine Learning for Network Routing. 2020 9th Mediterranean Conference on Embedded Computing (MECO). :1–1.
Though currently a “hot topic”, over the past fifteen years [1][2], there has been significant work on the use of machine learning to design large scale computer-communication networks, motivated by the complexity of the systems that are being considered and the unpredictability of their workloads. A topic of great concern has been security [3] and novel techniques for detecting network attacks have been developed based on Machine Learning [8]. However the main challenge with Machine Learning methods in networks has concerned their compatibility with the Internet Protocol and with legacy systems, and a major step forward has come from the establishment of Software Defined Networks (SDN) [4] which delegate network routing to specific SDN routers [4]. SDN has become an industry standard for concentrating network management and routing decisions within specific SDN routers that download the selected paths periodically to network routers, which operate otherwise under the IP protocol. In this paper we describe our work on real-time control of Security and Privacy [7], Energy Consumption and QoS [6] of packet networks using Machine Learning based on the Cognitive Packet Network [9] principles and their application to the H2020 SerIoT Project [5].
Sohail, Muhammad, Zheng, Quan, Rezaiefar, Zeinab, Khan, Muhammad Alamgeer, Ullah, Rizwan, Tan, Xiaobin, Yang, Jian, Yuan, Liu.  2020.  Triangle Area Based Multivariate Correlation Analysis for Detecting and Mitigating Cache Pollution Attacks in Named Data Networking. 2020 3rd International Conference on Hot Information-Centric Networking (HotICN). :114–121.
The key feature of NDN is in-network caching that every router has its cache to store data for future use, thus improve the usage of the network bandwidth and reduce the network latency. However, in-network caching increases the security risks - cache pollution attacks (CPA), which includes locality disruption (ruining the cache locality by sending random requests for unpopular contents to make them popular) and False Locality (introducing unpopular contents in the router's cache by sending requests for a set of unpopular contents). In this paper, we propose a machine learning method, named Triangle Area Based Multivariate Correlation Analysis (TAB-MCA) that detects the cache pollution attacks in NDN. This detection system has two parts, the triangle-area-based MCA technique, and the threshold-based anomaly detection technique. The TAB-MCA technique is used to extract hidden geometrical correlations between two distinct features for all possible permutations and the threshold-based anomaly detection technique. This technique helps our model to be able to distinguish attacks from legitimate traffic records without requiring prior knowledge. Our technique detects locality disruption, false locality, and combination of the two with high accuracy. Implementation of XC-topology, the proposed method shows high efficiency in mitigating these attacks. In comparison to other ML-methods, our proposed method has a low overhead cost in mitigating CPA as it doesn't require attackers' prior knowledge. Additionally, our method can also detect non-uniform attack distributions.
Marechal, Emeline, Donnet, Benoit.  2020.  Network Fingerprinting: Routers under Attack. 2020 IEEE European Symposium on Security and Privacy Workshops (EuroS PW). :594–599.
Nowadays, simple tools such as traceroute can be used by attackers to acquire topology knowledge remotely. Worse still, attackers can use a lightweight fingerprinting technique, based on traceroute and ping, to retrieve the routers brand, and use that knowledge to launch targeted attacks. In this paper, we show that the hardware ecosystem of network operators can greatly vary from one to another, with all potential security implications it brings. Indeed, depending on the autonomous system (AS), not all brands play the same role in terms of network connectivity. An attacker could find an interest in targeting a specific hardware vendor in a particular AS, if known defects are present in this hardware, and if the AS relies heavily on it for forwarding its traffic.
Kolomoitcev, V. S..  2020.  Effectiveness of Options for Designing a Pattern of Secure Access ‘Connecting Node’. 2020 Wave Electronics and its Application in Information and Telecommunication Systems (WECONF). :1–5.
The purpose of the work was to study the fault- tolerant pattern of secure access of computer system nodes to external network resources - the pattern of secure access `Connecting node'. The pattern of secure access `Connecting node' includes a group/cluster (or several groups) of routers, a computing node that includes hardware and software for information protection and communication channels that connect it to the end nodes of the computing system and the external network (network resources that are not controlled by the information protection system). The efficiency assessment and comparative analysis of options for designing a pattern of secure access `Connecting node' according to various efficiency criteria were carried out. In this work, an assessment of the individual and comprehensive efficiency index was carried out. It was assumed that the system is recoverable. The effectiveness of using some options of designing a pattern of secure access in terms of the operational availability factor, as well as a group of parameters - the operational availability factor, service delays of information protection system and the grade of information protection.
Adithyan, A., Nagendran, K., Chethana, R., Pandy D., Gokul, Prashanth K., Gowri.  2020.  Reverse Engineering and Backdooring Router Firmwares. 2020 6th International Conference on Advanced Computing and Communication Systems (ICACCS). :189–193.
Recently, there has been a dramatic increase in cyber attacks around the globe. Hundreds of 0day vulnerabilities on different platforms are discovered by security researchers worldwide. The attack vectors are becoming more and more difficult to be discovered by any anti threat detection engine. Inorder to bypass these smart detection mechanisms, attackers now started carrying out attacks at extremely low level where no threat inspection units are present. This makes the attack more stealthy with increased success rate and almost zero detection rate. A best case example for this scenario would be attacks like Meltdown and Spectre that targeted the modern processors to steal information by exploiting out-of-order execution feature in modern processors. These types of attacks are incredibly hard to detect and patch. Even if a patch is released, a wide range of normal audience are unaware of this both the vulnerability and the patch. This paper describes one such low level attacks that involves the process of reverse engineering firmwares and manually backdooring them with several linux utilities. Also, compromising a real world WiFi router with the manually backdoored firmware and attaining reverse shell from the router is discussed. The WiFi routers are almost everywhere especially in public places. Firmwares are responsible for controlling the routers. If the attacker manipulates the firmware and gains control over the firmware installed in the router, then the attacker can get a hold of the network and perform various MITM attacks inside the network with the help of the router.
Pimple, Nishant, Salunke, Tejashree, Pawar, Utkarsha, Sangoi, Janhavi.  2020.  Wireless Security — An Approach Towards Secured Wi-Fi Connectivity. 2020 6th International Conference on Advanced Computing and Communication Systems (ICACCS). :872–876.
In today's era, the probability of the wireless devices getting hacked has grown extensively. Due to the various WLAN vulnerabilities, hackers can break into the system. There is a lack of awareness among the people about security mechanisms. From the past experiences, the study reveals that router security encrypted protocol is often cracked using several ways like dictionary attack and brute force attack. The identified methods are costly, require extensive hardware, are not reliable and do not detect all the vulnerabilities of the system. This system aims to test all router protocols which are WEP, WPA, WPA2, WPS and detect the vulnerabilities of the system. Kali Linux version number 2.0 is being used over here and therefore the tools like airodump-ng, aircrack-ng are used to acquire access point pin which gives prevention methods for detected credulity and aims in testing various security protocols to make sure that there's no flaw which will be exploited.
Chinthavali, M., Starke, M., Moorthy, R..  2020.  An Intelligent Energy Router for Managing Behind-the-Meter Resources and Assets. 2020 IEEE Power Energy Society Innovative Smart Grid Technologies Conference (ISGT). :1–5.
With increase in distributed energy resources (DERs) and smart loads, each energy resource and load need a separate power conversion system leading to complex coordination and interaction, reduced energy conversion efficiency, coordinating compliance to grid standards (IEEE 1547) from multiple sources, reduced security. Also, multiple vendors with legacy system designs and proprietary communications interfaces result in redundancy and increase in cost of power electronics systems. This paper presents an energy router concept for buildings applications which provides autonomous power flow between sources and loads with a novel agent-based software interface.
Zhu, Fangzhou, Liu, Liang, Meng, Weizhi, Lv, Ting, Hu, Simin, Ye, Renjun.  2020.  SCAFFISD: A Scalable Framework for Fine-Grained Identification and Security Detection of Wireless Routers. 2020 IEEE 19th International Conference on Trust, Security and Privacy in Computing and Communications (TrustCom). :1194–1199.

The security of wireless network devices has received widespread attention, but most existing schemes cannot achieve fine-grained device identification. In practice, the security vulnerabilities of a device are heavily depending on its model and firmware version. Motivated by this issue, we propose a universal, extensible and device-independent framework called SCAFFISD, which can provide fine-grained identification of wireless routers. It can generate access rules to extract effective information from the router admin page automatically and perform quick scans for known device vulnerabilities. Meanwhile, SCAFFISD can identify rogue access points (APs) in combination with existing detection methods, with the purpose of performing a comprehensive security assessment of wireless networks. We implement the prototype of SCAFFISD and verify its effectiveness through security scans of actual products.

2021-03-29
Moti, Z., Hashemi, S., Jahromi, A. N..  2020.  A Deep Learning-based Malware Hunting Technique to Handle Imbalanced Data. 2020 17th International ISC Conference on Information Security and Cryptology (ISCISC). :48–53.
Nowadays, with the increasing use of computers and the Internet, more people are exposed to cyber-security dangers. According to antivirus companies, malware is one of the most common threats of using the Internet. Therefore, providing a practical solution is critical. Current methods use machine learning approaches to classify malware samples automatically. Despite the success of these approaches, the accuracy and efficiency of these techniques are still inadequate, especially for multiple class classification problems and imbalanced training data sets. To mitigate this problem, we use deep learning-based algorithms for classification and generation of new malware samples. Our model is based on the opcode sequences, which are given to the model without any pre-processing. Besides, we use a novel generative adversarial network to generate new opcode sequences for oversampling minority classes. Also, we propose the model that is a combination of Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) to classify malware samples. CNN is used to consider short-term dependency between features; while, LSTM is used to consider longer-term dependence. The experiment results show our method could classify malware to their corresponding family effectively. Our model achieves 98.99% validation accuracy.
Olaimat, M. Al, Lee, D., Kim, Y., Kim, J., Kim, J..  2020.  A Learning-based Data Augmentation for Network Anomaly Detection. 2020 29th International Conference on Computer Communications and Networks (ICCCN). :1–10.
While machine learning technologies have been remarkably advanced over the past several years, one of the fundamental requirements for the success of learning-based approaches would be the availability of high-quality data that thoroughly represent individual classes in a problem space. Unfortunately, it is not uncommon to observe a significant degree of class imbalance with only a few instances for minority classes in many datasets, including network traffic traces highly skewed toward a large number of normal connections while very small in quantity for attack instances. A well-known approach to addressing the class imbalance problem is data augmentation that generates synthetic instances belonging to minority classes. However, traditional statistical techniques may be limited since the extended data through statistical sampling should have the same density as original data instances with a minor degree of variation. This paper takes a learning-based approach to data augmentation to enable effective network anomaly detection. One of the critical challenges for the learning-based approach is the mode collapse problem resulting in a limited diversity of samples, which was also observed from our preliminary experimental result. To this end, we present a novel "Divide-Augment-Combine" (DAC) strategy, which groups the instances based on their characteristics and augments data on a group basis to represent a subset independently using a generative adversarial model. Our experimental results conducted with two recently collected public network datasets (UNSW-NB15 and IDS-2017) show that the proposed technique enhances performances up to 21.5% for identifying network anomalies.
Yilmaz, I., Masum, R., Siraj, A..  2020.  Addressing Imbalanced Data Problem with Generative Adversarial Network For Intrusion Detection. 2020 IEEE 21st International Conference on Information Reuse and Integration for Data Science (IRI). :25–30.

Machine learning techniques help to understand underlying patterns in datasets to develop defense mechanisms against cyber attacks. Multilayer Perceptron (MLP) technique is a machine learning technique used in detecting attack vs. benign data. However, it is difficult to construct any effective model when there are imbalances in the dataset that prevent proper classification of attack samples in data. In this research, we use UGR'16 dataset to conduct data wrangling initially. This technique helps to prepare a test set from the original dataset to train the neural network model effectively. We experimented with a series of inputs of varying sizes (i.e. 10000, 50000, 1 million) to observe the performance of the MLP neural network model with distribution of features over accuracy. Later, we use Generative Adversarial Network (GAN) model that produces samples of different attack labels (e.g. blacklist, anomaly spam, ssh scan) for balancing the dataset. These samples are generated based on data from the UGR'16 dataset. Further experiments with MLP neural network model shows that a balanced attack sample dataset, made possible with GAN, produces more accurate results than an imbalanced one.

Peng, Y., Fu, G., Luo, Y., Hu, J., Li, B., Yan, Q..  2020.  Detecting Adversarial Examples for Network Intrusion Detection System with GAN. 2020 IEEE 11th International Conference on Software Engineering and Service Science (ICSESS). :6–10.
With the increasing scale of network, attacks against network emerge one after another, and security problems become increasingly prominent. Network intrusion detection system is a widely used and effective security means at present. In addition, with the development of machine learning technology, various intelligent intrusion detection algorithms also start to sprout. By flexibly combining these intelligent methods with intrusion detection technology, the comprehensive performance of intrusion detection can be improved, but the vulnerability of machine learning model in the adversarial environment can not be ignored. In this paper, we study the defense problem of network intrusion detection system against adversarial samples. More specifically, we design a defense algorithm for NIDS against adversarial samples by using bidirectional generative adversarial network. The generator learns the data distribution of normal samples during training, which is an implicit model reflecting the normal data distribution. After training, the adversarial sample detection module calculates the reconstruction error and the discriminator matching error of sample. Then, the adversarial samples are removed, which improves the robustness and accuracy of NIDS in the adversarial environment.
Gupta, S., Buduru, A. B., Kumaraguru, P..  2020.  imdpGAN: Generating Private and Specific Data with Generative Adversarial Networks. 2020 Second IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (TPS-ISA). :64–72.
Generative Adversarial Network (GAN) and its variants have shown promising results in generating synthetic data. However, the issues with GANs are: (i) the learning happens around the training samples and the model often ends up remembering them, consequently, compromising the privacy of individual samples - this becomes a major concern when GANs are applied to training data including personally identifiable information, (ii) the randomness in generated data - there is no control over the specificity of generated samples. To address these issues, we propose imdpGAN-an information maximizing differentially private Generative Adversarial Network. It is an end-to-end framework that simultaneously achieves privacy protection and learns latent representations. With experiments on MNIST dataset, we show that imdpGAN preserves the privacy of the individual data point, and learns latent codes to control the specificity of the generated samples. We perform binary classification on digit pairs to show the utility versus privacy trade-off. The classification accuracy decreases as we increase privacy levels in the framework. We also experimentally show that the training process of imdpGAN is stable but experience a 10-fold time increase as compared with other GAN frameworks. Finally, we extend imdpGAN framework to CelebA dataset to show how the privacy and learned representations can be used to control the specificity of the output.