Lecturers
Prof. Eduard Axel Jorswieck (Technische Universitat Braunschweig, Germany)
Eduard Axel Jorswieck is the managing director of the Institute for Communications Technology and Full Professor at Technische Universit ̈at Braunschweig, Germany. From 2008 until 2019, he was the head of the Chair of Communications Theory and Full Professor at Dresden University of Technology, Germany. His main research interests are in the broad area of communications. He has published more than 150 journal papers, 15 book chapters, 3 monographs, and some 300 conference papers. Dr. Jorswieck is a Fellow of IEEE. Since 2017, he has been serving as Editor-in-Chief for the EURASIP Journal on Wireless Communications and Networking. Since 2021, he serves on the editorial board of IEEE Transactions on Communications. From 2011 to 2015, he acted as Associate Editor for IEEE Transactions on Signal Processing. Since 2008, continuing until 2011, he has served as an Associate Editor for IEEE Signal Processing Letters. From 2012 until 2013 he served as Senior Associate Editor for IEEE Signal Processing Letter. Since 2013, he serves as Editor for IEEE Transactions on Wireless Communications. Since 2016, he serves as Associate Editor for IEEE Transactions on Information Forensics and Security. In 2006, he received the IEEE Signal Processing Society Best Paper Award.
Prof. Babak Hossein Khalaj (Sharif University of Technology, Iran)
Babak Hossein Khalaj (Senior Member, IEEE) received his B.Sc. degree in Electrical Engineering from Sharif University of Technology, Tehran, Iran, in 1989, and M.Sc. and Ph.D. degrees in Electrical Engineering from Stanford University, Stanford, CA, USA, in 1993 and 1996, respectively. He is currently a Full Professor at Department of Electrical Engineering of Sharif University of Technology and the Director of Centre for Information Systems and Data Science at Sharif University. He has been with the pioneering team at Stanford University, where he was involved in adoption of multi-antenna arrays in mobile networks. Since 1999, he has been a Senior Consultant in the area of data communications, and from 2006 to 2007, a Visiting Professor with CEIT, San Sebastian, Spain. He has co-authored many papers in signal processing and digital communications and holds four U.S. patents. He was the recipient of the Alexander von Humboldt Fellowship from 2007 to 2008 and Nokia Visiting Professor Scholarship in 2018.
Mehdi Letafati (Sharif University of Technology, Iran)
Mehdi Letafati received his B.Sc and M.Sc. degrees in Electrical Engineering from Sharif University of Technology, Tehran, Iran, in 2019 and 2021, respectively. He is currently pursuing the Ph.D. degree in Electrical Engineering (communications systems) at Sharif University of Technology. He was a program attendee at Cornell, Maryland, Max-Planck Pre-Doctoral Research School, Saarbruecken, Germany, in August 2020. His research interests include both the theoretical and practical aspects of learning-based communication security, privacy in data science, and secure digital healthcare. He is also interested in the intersection of deep learning and information theory. He was honored to be ranked 4th among all participants in the Nationwide University Entrance Exam in 2015, and since then he has been a recipient of Iran’s National Elite Foundation’s scholarships. He was a recipient of the Exceptional Talent for outstanding performance during his undergraduate studies. Mehdi serves as a peer reviewer for top IEEE journals, including the IEEE Internet of Things Journal, IEEE Transactions on Signal Processing, and IEEE Wireless Communications Letters.
Abstract
Despite the development of different mechanisms to secure the core network of communication systems, the wireless edge of B5G and 6G systems is still vulnerable to security and privacy risks due to the inherent broadcast nature of wireless medium. To overcome this challenge, physical layer security (PLS) solutions have been envisioned to be leveraged for 6G networks thanks to their intrinsic capability of being adapted to the communication medium, providing agile security for different scenarios. As the 6G is envisioned to bring device-level intelligence, the capabilities of deep learning (DL) algorithms can be incorporated into PLS protocols, resulting in novel context-aware learning-based secure frameworks. In this tutorial, we provide a comprehensive overview on learning-based physical layer security (PLS) techniques as one of the key enablers for safeguarding the 6G wireless networks. As a preliminary, we first formulate the PLS framework, and review two main classes of PLS solutions, i.e., key-less and key-based PLS. Then, we address some of the state-of-the-art concepts and protocols in PLS, including “Quality-of-Security” (QoSec), multi-user mMIMO PHY key agreement, and man-in-the-middle (MitM) resilient key generation. In the next part of this tutorial, we take into account the context of communicated data, and focus on learning-based PLS to realize context-aware intelligent solutions against passive and active adversarial attacks. i) For the key-less PLS, we introduce different DL-based approaches for designing wiretap codes and enhancing the QoSec. We further introduce an end-to-end learning-based secure framework that privatizes sensitive data against adversarial neural networks. ii) For the key based PLS, recurrent-based neural networks and reservoir learning approach are addressed in the context of wireless key generation. Finally, we provide future directions, including the potential use of intelligent PLS solutions for the future e-health services to provide interested attendees with useful insights.
Motivation and Context
Security of the 5G standard relies upon conventional cryptographic approaches such as the elliptic curve cryptography (ECC) to fulfill the confidentiality and authentication requirements. Nevertheless, the modern era of 6G as the fabric that facilitates numerous peer-to-peer communications on-the-fly, undermines the performance of conventional security solutions. In addition, the resource-limited nodes, e.g. the e-health on- and in-body nodes, make the implementation of pre-deployed keys impractical. Therefore, a paradigm shift from conventional (all-purpose) complexity-based security techniques towards application-centric, context-aware, and intelligent security solutions is highly required to enhance the “quality of security”. Besides, safeguarding the “connected intelligence” in 6G requires intelligent data-driven approaches to guarantee security, robustness, and resiliency against adversarial threats. PHY layer security (PLS) has been envisioned as a promising framework to migrate from the conventional complexity-based solutions towards lightweight context-aware techniques for 6G networks. PLS protocols are thoroughly decentralized and do not rely on any specific third-party infrastructure. In addition to rendering flexible secure frameworks against passive and active adversarial attacks, PLS also achieves information-theoretic security guarantees and thereby post-quantum security. As the communication networks are facing a new trend of transferring functionalities from PHY to higher layers by employing software-centric solutions, we show in this tutorial that this key idea can be incorporated into the PLS protocols as well. That is, we address how we can leverage context-aware learning algorithms to improve the secrecy performance. The integration of machine learning (ML) and DL methods with novel security mechanisms makes it highly attractive for leading AI companies to develop Security-as-a-Service (SecaaS) solutions. Therefore, learningbased PLS solutions proposed in this tutorial will shed light on further developments of SecaaS products.
The objective of this tutorial is to introduce a modern concept for realizing context-aware intelligent security for the future 6G wireless systems, which includes learning-based PLS techniques. This tutorial also aims to provide attendees with insightful concepts and material regarding context-aware wireless security, including QoSec, neural wiretap codes, secure neural encoding, MitM-resiliency, mMIMO multi-user key agreement, and end-to-end DL for wireless security.
Structure and Content
In this tutorial, learning-based physical layer security (PLS) techniques are introduced as one of the key enabler for safeguarding the sixth generation (6G) wireless networks in a context-aware manner. As a preliminary, we first formulate the PLS framework, and review two main classes of PLS solutions, i.e., key-less and key-based PLS. Then, we address some of the state-of-the-art concepts and protocols in PLS, including “Quality-of-Security” (QoSec), man-in-the-middle (MitM) resiliency, multi-user massive MIMO (mMIMO) PHY key agreement schemes, etc., which are realized via novel PLS-based schemes. In the next part of the tutorial, we take into account the context of data communicated through the network, and focus on learning-based PLS, aiming to realize context-aware intelligent solutions against passive and active adversarial attacks. For the key-less PLS we introduce different deep learning-based approaches for i) enhancing the QoSec, ii) designing wiretap codes, and iii) realizing an end-to-end secure framework that privatizes sensitive data against adversarial neural networks. For the key-based PLS, we address recurrent-based neural networks and reservoir learning as two approaches to design context-aware security solutions. Finally, we address future directions, including the potential use of intelligent PLS solutions for the future e-health services to provide interested attendees with useful insights. Most of the topics that will be covered in this tutorial are based on the recent publications of the instructors in top IEEE journals and conferences.
An outline of the tutorial, including a tentative time schedule:
The first part of the tutorial (~ 1:15’):
~ 35 min.:
- Motivation for PLS: Importance of PHY security for 6G communication networks
- Introduction to PLS solutions in the context of wireless communications
-
- Model and problem Formulation
- Key-less and key-based techniques to ensure PLS
- Active and passive attacks
~ 40 min.:
- Introduction to Secret Key Generation
- Physical layer Secret Key Generation
-
- Per user key generation in multi-user massive MIMO
- MitM attack resistant SKG for multi-carrier MIMO systems
The second part of the tutorial (~ 1:30’):
~ 15 min.:
- Introduction to deep learning: An overview of different types of neural networks
~ 1 hour:
- Leveraging deep learning in PLS: Towards context-aware intelligence for wireless security
-
- Deep learning-aided key-less PLS
- Learning-aided transmission design to maximize QoSec
- Neural network wiretap autoencoder
- End-to-end secure image transmission: Neural encoder/decoder to privatize data against multiple adversarial networks
- Deep learning for key agreement schemes under active attacks
- Reservoir learning for PLS: A lightweight training mechanism
~ 15 min.:
- Future Directions and Suggested Works