|
BGK 596E Seminars
CRN 12338
|
- This page is updated weekly.
- You may stay tuned.
|
| 7 |
| Title |
Measurements of Security Metrics for Wireless Communications |
| Speaker |
Fatih Keskin |
| Date |
December 12th, 2025 |
| Time |
10:30 (GMT+3) |
| Microsoft Teams |
Meeting ID:
***
Passcode:
***
(Please, contact the secretary's office.)
|
Abstract
Security metrics have become increasingly central to the design, development, and operation of wireless networks and communication systems. Although widely used, most existing studies address only limited aspects of these metrics or focus on narrowly defined objectives. Security metrics offer both qualitative and quantitative insights into how resilient a wireless system is, and they play a crucial role in evaluating security posture and achieving security objectives. However, despite the availability of multiple assessment methods, there is still no comprehensive and structured classification of metrics specifically designed for wireless or mobile communication environments. This study aims to bridge that gap by presenting an organized overview of commonly adopted security indicators within wireless communication systems. To achieve this, we group current assessment approaches into two main categories: network-centric and host-centric evaluations. Host-oriented metrics are further divided into probabilistic and non-probabilistic models, while network-focused metrics are classified into route-based and non-route-based measures. The classification framework proposed in this research provides a fresh perspective for analyzing the security level of wireless networks and communication infrastructures. By contributing a more systematic structure for assessing risk and resilience, this study supports better policy-making, improved practical deployments, and a stronger foundation for ensuring secure and adaptable wireless systems.
|
|
| 6 |
| Title |
Secure Aggregation for Privacy-preserving Federated Learning in Vehicular Networks |
| Speaker |
Zeynep Eda Ustalar |
| Date |
December 5th, 2025 |
| Time |
10:30 (GMT+3) |
| Microsoft Teams |
Meeting ID:
***
Passcode:
***
(Please, contact the secretary's office.)
|
Abstract
The increase in interest in autonomous driving systems has reinforced the importance of ensuring privacy and security in ITS. This paper introduces a new secure aggregation protocol to ensure privacy requirements during Federated Learning in vehicular network communications. According to the presented protocol structure, each of the SCMS and FL PKI components has their own key in the communication flow. Encryption, decryption and signature in each round are handled with their own keys. So, each component can see only necessary information. This approach aims to downgrade insider attacks and ensure the privacy of vehicle information. The proposed protocol structure is analyzed under headings such as private data protection, malicious behavior reporting, independent trust domain logic etc. In addition to this security analysis, calculation, communication and storage costs on vehicle and server basis are evaluated. Experimental results show that although there exist challenges and limitations like connection stability while driving and resource scalability, introduced protocol ensures and maintains the essential data privacy in federated learning in vehicular networks.
|
|
| 5 |
| Title |
Decentralized Vulnerability Disclosure Using Permissioned Blockchain: A Secure and Transparent Alternative to Centralized CVE Management |
| Speaker |
Novruz Amirov |
| Date |
November 28th, 2025 |
| Time |
10:30 (GMT+3) |
| Microsoft Teams |
Meeting ID:
***
Passcode:
***
(Please, contact the secretary's office.)
|
Abstract
This study introduces a decentralized, blockchain-driven framework for publishing Common Vulnerabilities and Exposures (CVEs) as an alternative to the predominantly centralized model managed by MITRE. The proposed solution employs a permissioned blockchain that restricts write privileges to authenticated CVE Numbering Authorities (CNAs) while maintaining public transparency. Through the integration of smart contracts, the framework enables essential functions such as embargoed vulnerability disclosures and decentralized governance. We assess the proposed approach against current practices, demonstrating its improvements in transparency, trust distribution, and auditability. Furthermore, we present a prototype built on Hyperledger Fabric to validate the feasibility of the model and discuss its potential impact on the evolution of vulnerability disclosure mechanisms.
|
|
| 4 |
| Title |
Artificial Intelligence and Quantum Cryptography |
| Speaker |
Elanur Tığlı |
| Date |
November 21st, 2025 |
| Time |
10:30 (GMT+3) |
| Microsoft Teams |
Meeting ID:
***
Passcode:
***
(Please, contact the secretary's office.)
|
Abstract
The rapid developments in artificial intelligence (AI) and quantum computing have significantly influenced modern cryptographic techniques. Particularly, AI-based methods have shown strong potential in enhancing the efficiency and resilience of cryptographic systems. However, the advent of quantum computers introduces a “quantum threat” that challenges existing security algorithms. This paper proposes a novel cryptographic approach called AI-assisted quantum cryptography, which leverages neural network models to optimize key distribution and threat detection processes. Unlike traditional cryptographic systems, AI-assisted quantum cryptography operates as an adaptive many-to-many framework, dynamically adjusting security parameters in response to detected quantum attacks. The proposed scheme integrates AI algorithms with quantum key distribution (QKD) protocols, enhancing both the speed and robustness of secure communications.
The basic model demonstrates effectiveness in controlled synchronous environments, while asynchronous implementations are also introduced for continuous and multi- session key management. AI-assisted quantum cryptography can efficiently maintain secure communication channels, even under evolving quantum threats, and serves as a foundation for future secure digital infrastructures. Potential vulnerabilities and areas for improvement will be addressed in a subsequent study titled Adaptive AI-Based Quantum Cryptography with Multi-Session Security.
|
|
| 3 |
| Title |
Encrypted Network Traffic Classification using Self-supervised Learning |
| Speaker |
Emre Can Tüzer |
| Date |
November 14th, 2025 |
| Time |
10:30 (GMT+3) |
| Microsoft Teams |
Meeting ID:
***
Passcode:
***
(Please, contact the secretary's office.)
|
Abstract
This paper proposes a novel two-stage self-supervised deep learning approach for encrypted network traffic classification that achieves high accuracy with only a small volume of labeled data, Critical applications including resource management, malware detection, and network provisioning are supported by network traffic classification. However, this categorization task has become difficult due to the widespread use of encrypted protocols in modern networks. Large volumes of labeled data are necessary for high accuracy in both traditional and deep learning-based supervised algorithms, but acquiring such data in actual network systems is costly, time-consuming, and requires skill. This research suggests a self-supervised learning method for classifying encrypted network traffic as a solution to this problem. With a limited amount of labeled data, the suggested approach seeks to get high accuracy. The efficacy of this method is confirmed by experiments conducted on three publicly accessible datasets: Orange'20, ISCX VPN-NonVPN, and QUIC. According to experimental results, the suggested approach performs about 3% more accurately than current state-of-the-art baseline approaches.
Furthermore, the model's transferability is remarkable; its accuracy drops by only 3% even after being pre-trained on a different dataset and refined using labeled data. For adaptive network monitoring in dynamic network conditions, this transferability is essential. These findings show that the analysis of encrypted traffic can be greatly enhanced by self-supervised representation learning, which also lessens the need for human-labeled data.
|
|
| 2 |
| Title |
LLM Agents Can Autonomously Hack Websites |
| Speaker |
Can Nair |
| Date |
November 7th, 2025 |
| Time |
10:30 (GMT+3) |
| Microsoft Teams |
Meeting ID:
***
Passcode:
***
(Please, contact the secretary's office.)
|
Abstract
In recent years, large language models (LLMs) have evolved to function autonomously as agents, capable of interacting with tools, reading documents, and planning multi-step tasks. This presentation covers the paper “LLM Agents Can Autonomously Hack Websites”, which investigates how LLM-based agents, under controlled experimental conditions, can autonomously perform tasks such as web reconnaissance, vulnerability discovery, and automated exploitation attempts. The experiments were conducted in a secure lab environment, following ethical guidelines, to highlight potential misuse without encouraging real-world attacks. The study shows that GPT-4-based agents can execute these tasks without prior knowledge of the vulnerabilities, demonstrating the offensive capabilities of autonomous LLM agents. These findings emphasize the dual-use potential of AI in cybersecurity and the need for careful consideration by researchers and policymakers.
|
|
| 1 |
| Title |
Approximation and Heuristic Approaches for the Metric Traveling Salesperson Problem (MTSP) |
| Speaker |
Burcu Sultan Orhan |
| Date |
October 31rd, 2025 |
| Time |
10:30 (GMT+3) |
| Microsoft Teams |
Meeting ID:
***
Passcode:
***
(Please, contact the secretary's office.)
|
Abstract
The Traveling Salesperson Problem (TSP) is a well-known NP-hard problem in combinatorial optimization that aims to determine the shortest possible tour visiting each city exactly once and returning to the starting point. The metric version of TSP (MTSP) assumes that the triangle inequality holds for the distance metric, i.e., d(i,j) + d(j,k) ≥ d(i,k). This assumption significantly influences the design and performance of approximation algorithms. This presentation reviews theoretical foundations and key approximation approaches for the MTSP, including the Christofides algorithm and later heuristic extensions. The focus is placed on algorithmic guarantees, complexity analysis, and the trade-off between theoretical optimality and practical efficiency.
|
|