Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 15 of 15 for “"Defense method"”.

  1. Generative Neural Network-Based Defense Methods Against Cyberattacks for Connected and Autonomous Vehicles

    … a generative adversarial network (GAN)-based defense method for the traffic sign classification system within the perception module of autonomous vehicles. The author evaluated the hybrid quantum-classical RBM-based intrusion detection framework on three separate real-world Fuzzy attack …

    clemson Repository record for Generative Neural Network-Based Defense Methods Against Cyberattacks for Connected and Autonomous Vehicles (opens in a new tab)

  2. Improving adversarial robustness of machine learning models

    … Neural Networks (DNNs) and aim to build robust defense methods. We identity three important factors of DNNs, including training data, representation learnings, and model deployment. We develop RanTrain, CDAD, and RAD respectively. RanTrain is a defense method which augments training data by …

    adelaide Repository record for Improving adversarial robustness of machine learning models (opens in a new tab)

  3. Security and Reliability of Vehicular Networks and Autonomous Vehicle Applications Using Artificial Intelligence and Edge Computing in a Cyber-Physical Systems Environment

    … In this dissertation, the author demonstrates methods and models to improve the security and reliability of CAV communication and applications under different scenarios using state-of-the-art technologies, such as software-defined networking, artificial intelligence, and edge computing. This …

    clemson Repository record for Security and Reliability of Vehicular Networks and Autonomous Vehicle Applications Using Artificial Intelligence and Edge Computing in a Cyber-Physical Systems Environment (opens in a new tab)

  4. A COMPACT NETWORK MODEL FOR LEARNING IN DISTRIBUTION SPACE

    Despite the superior performance of deep learning methods in various tasks, challenges remain in the area of regression on function spaces. In this thesis, we address the problem of regression where the data are distributions. However, neural networks are not designed for distribution inputs - the …

    nus Repository record for A COMPACT NETWORK MODEL FOR LEARNING IN DISTRIBUTION SPACE (opens in a new tab) Download file for A COMPACT NETWORK MODEL FOR LEARNING IN DISTRIBUTION SPACE (opens in a new tab)

  5. The Impact of Parasitic DC And AC Sources on the Security of the KLJN Secure Key Exchange Scheme

    … the usual current/voltage comparison-based defense method that exposes active attacks or parasitic features (such as wire resistance based information leaks) does not workhere. The attack is successfully demonstrated and we proposed defense methods against the attack are shown. The second …

    tamu Repository record for The Impact of Parasitic DC And AC Sources on the Security of the KLJN Secure Key Exchange Scheme (opens in a new tab)

  6. Improving additional adversarial robustness for classification

    … for designing adversarial examples and methods to defending against them. This paper consists of two projects. In our first work, we propose an approach by leveraging cognitive salience to enhance additional robustness on top of these methods. Specifically, for image classification, we …

    wustl Repository record for Improving additional adversarial robustness for classification (opens in a new tab)

  7. Reliable and structural deep neural networks

    … part of this dissertation proposed an effective defense method against adversarial examples. We introduced an ensemble generative network with feedback loops, which use the feature-level denoising modules to improve the defense capability for adversarial examples. We then discussed the …

    missouri Repository record for Reliable and structural deep neural networks (opens in a new tab)

  8. Secure and Robust deep learning in computer vision

    … in computer vision. First, we propose a new defense method called Multi-Pronged Defense (MPD), which protects deep neural networks from backdoor attacks—hidden manipulations that cause models to behave incorrectly when triggered. MPD combines semi-supervised learning, balanced data sampling, …

    mcmaster Repository record for Secure and Robust deep learning in computer vision (opens in a new tab)

  9. An Exposition of User Privacy on the Web: Fingerprinting and Tracking

    … attacks to breach privacy, (2) design of privacy defense tools, (3) measurement of privacy compliance, and (4) measurement of browser fingerprinting in targeting and tracking. The first part involves the de-velopment of an attack called “Gummy Browser” that targets fingerprint-based tracking …

    tamu Repository record for An Exposition of User Privacy on the Web: Fingerprinting and Tracking (opens in a new tab)

  10. Overlay Networks and Distributed Denial of Service Attacks: Overview, study and evaluation of an application-enabled approach.

    … resource provides to the victim. The most common method to perpetrate DDoS attack is flooding the network with malicious packets to exhaust the network resources. This work is based on the fact that many network-based applications commonly open some known port(s) to communicate with their users; …

    chalmers Repository record for Overlay Networks and Distributed Denial of Service Attacks: Overview, study and evaluation of an application-enabled approach. (opens in a new tab)

  11. Securing Connected and Automated Surveillance Systems Against Network Intrusions and Adversarial Attacks

    … cameras. The first solution is a novel method for network flow features transformation, named TempoCode. It introduces a temporal codebook-based encoding of flow features based on capturing the key patterns of benign traffic in a learnt temporal codebook. The second solution takes an …

    ottawa-retro Repository record for Securing Connected and Automated Surveillance Systems Against Network Intrusions and Adversarial Attacks (opens in a new tab)

  12. Backdoor in AI: Algorithms, Attacks, and Defenses

    … we can then proceed to develop advanced defense algorithms. First, for exploring the new backdoor attack surface, we propose a training-free backdoor attack approach which is different from the traditional backdoor insertion method where backdoor behaviors are injected by training the …

    rice Repository record for Backdoor in AI: Algorithms, Attacks, and Defenses (opens in a new tab)

  13. Cyber-Secure and Resilient Networked Cyber-Physical Systems

    … then introduce a general-purpose cyber-attack defense method called Dynamic Watermarking (DW) to ensure cyber-secure and safe operations against potential cyber-attacks or cyber-intrusions. The DW method requires a model of the plant, which may not be available in some systems; for example, if …

    tamu Repository record for Cyber-Secure and Resilient Networked Cyber-Physical Systems (opens in a new tab)

  14. Exploration of New Passive and Active Attacks and Defense Methods for the KLJN and the VMG-KLJN Secure Key Exchangers

    … foundation of security. However, the VMG-KLJN method can, with appropriate countermeasures, be sufficiently secure for practical situations. Our first study reveals that under practical conditions with nonzero cable capacitance and inductance, the VMG-KLJN scheme is vulnerable to certain …

    tamu Repository record for Exploration of New Passive and Active Attacks and Defense Methods for the KLJN and the VMG-KLJN Secure Key Exchangers (opens in a new tab)

  15. Towards Personalized, Privacy-preserving, and Explainable AI for Healthcare

    … networks. Additionally, we suggest a robust defense method against gradient inversion attacks on medical image data via latent space data perturbation and minimax optimization in Chapter 5.</p><p>We have summarized the significance and unresolved challenges in remote health monitoring. …

    duke Repository record for Towards Personalized, Privacy-preserving, and Explainable AI for Healthcare (opens in a new tab) Embargoed until 2027-01-13