2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)最新文献

{"title":"Beta Poisoning Attacks Against Machine Learning Models: Extensions, Limitations and Defenses","authors":"Atakan Kara, Nursena Koprucu, M. E. Gursoy","doi":"10.1109/TPS-ISA56441.2022.00031","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00031","url":null,"abstract":"The rise of machine learning (ML) has made ML models lucrative targets for adversarial attacks. One of these attacks is Beta Poisoning, which is a recently proposed training-time attack based on heuristic poisoning of the training dataset. While Beta Poisoning was shown to be effective against linear ML models, it was originally developed with a fixed Gaussian Kernel Density Estimator (KDE) for likelihood estimation, and its effectiveness against more advanced, non-linear ML models has not been explored. In this paper, we advance the state of the art in Beta Poisoning attacks by making three novel contributions. First, we extend the attack so that it can be executed with arbitrary KDEs and norm functions. We integrate Gaussian, Laplacian, Epanechnikov and Logistic KDEs with three norm functions, and show that the choice of KDE can significantly impact attack effectiveness, especially when attacking linear models. Second, we empirically show that Beta Poisoning attacks are ineffective against non-linear ML models (such as neural networks and multi-layer perceptrons), even with our extensions. Results imply that the effectiveness of the attack decreases as model non-linearity and complexity increase. Finally, our third contribution is the development of a discriminator-based defense against Beta Poisoning attacks. Results show that our defense strategy achieves 99% and 93% accuracy in identifying poisoning samples on MNIST and CIFAR-10 datasets, respectively.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"245 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115626203","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ML-FEED: Machine Learning Framework for Efficient Exploit Detection","authors":"Tanujay Saha, Tamjid Al-Rahat, N. Aaraj, Yuan Tian, N. Jha","doi":"10.1109/TPS-ISA56441.2022.00027","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00027","url":null,"abstract":"Machine learning (ML)-based methods have recently become attractive for detecting security vulnerability exploits. Unfortunately, state-of-the-art ML models like long short-term memories (LSTMs) and transformers incur significant computation overheads. This overhead makes it infeasible to deploy them in real-time environments. We propose a novel ML-based exploit detection model, ML-FEED, that enables highly efficient inference without sacrificing performance. We develop a novel automated technique to extract vulnerability patterns from the Common Weakness Enumeration (CWE) and Common Vulnerabilities and Exposures (CVE) databases. This feature enables ML-FEED to be aware of the latest cyber weaknesses. Second, it is not based on the traditional approach of classifying sequences of application programming interface (API) calls into exploit categories. Such traditional methods that process entire sequences incur huge computational overheads. Instead, ML-FEED operates at a finer granularity and predicts the exploits triggered by every API call of the program trace. Then, it uses a state table to update the states of these potential exploits and track the progress of potential exploit chains. ML-FEED also employs a feature engineering approach that uses natural language processing-based word embeddings, frequency vectors, and one-hot encoding to detect semantically-similar instruction calls. Then, it updates the states of the predicted exploit categories and triggers an alarm when a vulnerability fingerprint executes. Our experiments show that ML-FEED is 72.9× and 75, 828.9× faster than state-of-the-art lightweight LSTM and transformer models, respectively. We trained and tested ML-FEED on 79 real-world exploit categories. It predicts categories of exploit in real-time with 98.2% precision, 97.4% recall, and 97.8% F1 score. These results also outperform the LSTM and transformer baselines. In addition, we evaluated ML-FEED on the attack traces of CVE vulnerability exploits in three popular Java libraries and detected all three reported critical vulnerabilities in them.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122412153","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Accountable AI for Healthcare IoT Systems","authors":"Prachi Bagave, Marcus Westberg, R. Dobbe, M. Janssen, A. Ding","doi":"10.1109/TPS-ISA56441.2022.00013","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00013","url":null,"abstract":"Various AI systems have taken a unique space in our daily lives, helping us in decision-making in critical as well as non-critical scenarios. Although these systems are widely adopted across different sectors, they have not been used to their full potential in critical domains such as the healthcare sector enabled by the Internet of Things (IoT). One of the important hindering factors for adoption is the implication for accountability of decisions and outcomes affected by an AI system, where the term accountability is understood as a means to ensure the performance of a system. However, this term is often interpreted differently in various sectors. Since the EU GDPR regulations and the US congress have emphasised the importance of enabling accountability in AI systems, there is a strong demand to understand and conceptualise this term. It is crucial to address various aspects integrated with accountability and understand how it affects the adoption of AI systems. In this paper, we conceptualise these factors affecting accountability and how it contributes to a trustworthy healthcare AI system. By focusing on healthcare IoT systems, our conceptual mapping will help the readers understand what system aspects those factors are contributing to and how they affect the system trustworthiness. Besides illustrating accountability in detail, we also share our vision towards causal interpretability as a means to enhance accountability for healthcare AI systems. The insights of this paper shall contribute to the knowledge of academic research on accountability, and benefit AI developers and practitioners in the healthcare sector.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133222374","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhanced Scanning in SDN Networks and its Detection using Machine Learning","authors":"Abdullah H. Alqahtani, John A. Clark","doi":"10.1109/TPS-ISA56441.2022.00032","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00032","url":null,"abstract":"Software-Defined Networking (SDN) is a networking approach that is dynamic and programmable, making network configuration easier and improving network efficiency. The separation of the control plane from the network plane and the global visibility of the controller to the whole network make the monitoring and collection of data much easier than in traditional networks. Advanced Persistent Threats (APTs) are notoriously hard to detect and prevent as they have sophisticated characteristics compared to traditional attacks. Little research has been carried out on the detection of APTs in the context of SDNs. In SDN, scanning is a fundamental part of the reconstruction of flow rules maintained at nodes (and underpins many further attacks). In this paper, we propose a more stealthy means of scanning within SDN networks, typical of the \"low and slow\" approach taken by APTs, and enhance a network scanning tool to implement it. We evaluate how well Machine Learning (ML) algorithms can detect such APT scanning activities inside SDN. We use the XGBoost classifier for the proposed detection model, achieving at least 97.8% in Accuracy, Recall, Precision and F1-measures using just 5 features. Datasets over different network sizes are generated to form the basis for experiments and are offered free public use.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129426199","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Developing Self-evolving Deepfake Detectors Against AI Attacks","authors":"Ian Miller, Dan Lin","doi":"10.1109/TPS-ISA56441.2022.00016","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00016","url":null,"abstract":"As deep-learning based image and video manipulation technology advances, the future of truth and information looks bleak. In particular, Deepfakes, wherein a person’s face can be transferred onto the face of someone else, pose a serious threat for potential spread of convincing misinformation that is drastic and ubiquitous enough to have catastrophic real-world consequences. To prevent this, an effective detection tool for manipulated media is needed. However, the detector cannot just be good, it has to evolve with the technology to keep pace with or even outpace the enemy. At the same time, it must defend against different attack types to which deep learning systems are vulnerable. To that end, in this paper, we review various methods of both attack and defense on AI systems, as well as modes of evolution for such a system. Then, we put forward a potential system that combines the latest technologies in multiple areas as well as several novel ideas to create a detection algorithm that is robust against many attacks and can learn over time with unprecedented effectiveness and efficiency.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130004111","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Managing Reputation Scores in a Blockchain-based Decentralized Marketplace","authors":"R. Mukkamala, S. Olariu, M. Aljohani, Sandeep Kalari","doi":"10.1109/TPS-ISA56441.2022.00020","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00020","url":null,"abstract":"The goal of a trust and reputation system is to provide buyers with a robust framework that allows them to select a seller based on accumulated evidence of the seller’s past behavior in the marketplace. The quality of such a trust and reputation system depends on the quality of the feedback provided by buyers. Being a subjective measure, the quality of buyer feedback is notoriously hard to assess. When feedback is provided by buyers from around the world, who may value different aspects of the same transaction differently, it is very hard to know when a buyer provides truthful feedback.In this work we assume a blockchain-based decentralized marketplace where a Smart Contract is associated with each transaction. In an attempt to reduce the uncertainty associated with buyer feedback, the novelty that we introduce in this work is that at the end of the transaction, the Smart Contract is responsible for providing feedback, replacing notoriously unreliable buyer feedback by a more objective assessment of how well the buyer and the seller have fulfilled their contractual obligations towards each other. While Smart Contracts are naturally associated with transactions, to the best of our knowledge, this is the first time Smart Contracts are empowered with providing transaction feedback.Our second main contribution is to show how to manage efficiently reputation scores in such a blockchain-based marketplace.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132814259","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Weaponizing IoT Sensors: When Table Choice Poses a Security Vulnerability","authors":"Gustavo Casqueiro, Sayed Erfan Arefin, T. Ashrafi, Abdul Serwadda, Hassan Wasswa","doi":"10.1109/TPS-ISA56441.2022.00029","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00029","url":null,"abstract":"The security threat posed by keyloggers on laptop and desktop computers is traditionally understood from the perspective of malware that directly reads keystrokes on the victim’s machine. While recent research on smart phone platforms has shown that motion/vibration sensors inbuilt in these phones also pose a keylogging threat, this line of attack has never been investigated in desktop and laptop settings given that no such sensors exist in these settings. In this paper, we show that the vibration dynamics of commonly used computer table materials transmit keyboard vibrations during typing with such fine granularity that keyboard typing locations (and hence keystrokes) could be learned from the vibrations. In practice such an attack would be executed by methodically rigging the underside of a computer table or keyboard itself with a series of motion sensors, and then mining the data generated by these sensors during typing. Taking the case of typical computer table materials such as glass, plastic, metal and wood, we study this line of attack and highlight scenarios where it poses a potent threat. Thanks to fast growing IoT platforms making available easy-to-use, fully featured, cheap sensors, we argue that this line of attack is accessible to even casual \"computer hackers\" having no knowledge of low-level hardware programming. The paper brings to light a previously unexplored privacy threat that security practitioners and end-users need to pay attention to as IoT goes mainstream.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115649738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enabling Location Based Services with Privacy and Integrity Protection in Untrusted Environments through Blockchain and Secure Computation","authors":"M. Amoretti, Alexandru Budianu, G. Caparra, Felice D’Agruma, Davide Ferrari, Gabriele Penzotti, L. Veltri, F. Zanichelli","doi":"10.1109/TPS-ISA56441.2022.00024","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00024","url":null,"abstract":"Privacy and integrity preservation of user data is a major challenge in the context of location based services, as the assumption of trusted relationship between the user and the service provider might be too strong. The question is: how to securely collect, store and process position, navigation and timing (PNT) information and/or georeferenced data, assuming that the service provider cannot be trusted? In this work, we propose an architecture that enables LBS with privacy and integrity in untrusted environments, leveraging blockchain and secure computation. We provide mechanisms for sharing and processing PNT information and/or georeferenced data, with a detailed description of the employed cryptographic schemes and algorithms. Furthermore, we provide a validation of the proposed architecture by means of an emulation-based testbed.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114280138","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ProProv: A Language and Graphical Tool for Specifying Data Provenance Policies","authors":"Kevin Dennis, Shamaria Engram, Tyler Kaczmarek, Jay Ligatti","doi":"10.1109/TPS-ISA56441.2022.00040","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00040","url":null,"abstract":"The Function-as-a-Service cloud computing paradigm has made large-scale application development convenient and efficient as developers no longer need to deploy or manage the necessary infrastructure themselves. However, as a consequence of this abstraction, developers lose insight into how their code is executed and data is processed. Cloud providers currently offer little to no assurance of the integrity of customer data. One approach to robust data integrity verification is the analysis of data provenance—logs that describe the causal history of data, applications, users, and non-person entities. This paper introduces ProProv, a new domain-specific language and graphical user interface for specifying policies over provenance metadata to automate provenance analyses.To evaluate the convenience and usability of the new ProProv interface, 61 individuals were recruited to construct provenance policies using both ProProv and the popular, general-purpose policy specification language Rego—used as a baseline for comparison. We found that, compared to Rego, the ProProv interface greatly increased the number of policies successfully constructed, improved the time taken to construct those policies, and reduced the failed-attempt rate. Participants successfully constructed 73% of the requested policies using ProProv, compared to 41% using Rego. To further evaluate the usability of the tools, participants were given a 10-question questionnaire measured using the System Usability Scale (SUS). The median SUS score for the graphical ProProv interface was above average and fell into the “excellent” category, compared to below average and “OK” for Rego. These results highlight the impacts that graphical domain-specific tools can have on the accuracy and speed of policy construction.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"8 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114088705","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Maintaining Review Credibility Using NLP, Reputation, and Blockchain","authors":"Zachary Zaccagni, R. Dantu, Kirill Morozov","doi":"10.1109/TPS-ISA56441.2022.00018","DOIUrl":"https://doi.org/10.1109/TPS-ISA56441.2022.00018","url":null,"abstract":"This paper presents a novel approach to review credibility in a marketplace, which leverages trust in reviews and reputation of the parties who provide them. We propose an architecture for a reputation-based review evaluation system, which is built on top of the blockchain system, in order to ensure correct and trustworthy assessments. In our proposal, trustworthiness of reviews is evaluated using NLP—specifically, sentimental analysis—and the reviewers’ reputations are adjusted according to this evaluation. These reputations are stored on the blockchain and used as an asset for the consensus process when mapped to stake. We introduce a new type of transaction, a review transaction, which stores the review data and evaluation results. In testing, our simulation results showed that the NLP component incurred a reasonable delay this new type of review transactions. Additionally, we measured the time required to add a standard payment transaction in Algorand and that for our review transaction and observed comparable results. Also, we observed that the NLP component ensures an accurate credible evaluation (compared to the ground truth) of the product review texts. With this new model, we have moved towards showing how NLP can used for self-regulating trust management in a decentralized marketplace ecosystem.","PeriodicalId":427887,"journal":{"name":"2022 IEEE 4th International Conference on Trust, Privacy and Security in Intelligent Systems, and Applications (TPS-ISA)","volume":"os7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128322373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}