Title: Open Set Wireless Transmitter Authorization: Deep Learning Approaches and Practical Considerations

Date and Time: November 19, 2021 at 11AM ET

Registration Process: Please register at https://tinyurl.com/wh89zv5w

Abstract: As the Internet of Things (IoT) continues to grow, ensuring the security of systems that rely on wireless IoT devices has become critically important. Deep learning-based passive physical layer transmitter authorization systems, referred as RF fingerprinting, have been introduced recently for this purpose. RF fingerprinting approaches use wireless signals directly to verify identity of radio frequency transmitters based on imperfections in their radio hardware. However, most of the existing work using machine learning for RF fingerprinting has mainly focused on classification approaches assuming a closed set of transmitters. In practice, most serious security problems would arise if malicious transmitters outside this closed set are misclassified and authorized. In this talk, we formulate the problem of recognizing authorized transmitters and rejecting new transmitters as open set recognition and anomaly detection. We consider approaches based on one and several binary classifiers, multi-class classifiers, and signal reconstruction, and study how these approaches scale with the number of authorized transmitters. We propose using a known set of unauthorized transmitters to assist the training and study its impact. The authorization robustness against temporal changes in fingerprints is also evaluated as a function of the approach and the dataset structure. For this work, we have created a large Wi-Fi dataset consisting of about 10 million packets sent by 174 off-the-shelf Wi-Fi radios and simultaneously captured by 41 USRPs during 4 captures performed in ORBIT testbed along a month. We have also developed generative deep learning methods to emulate unauthorized signal samples for the augmentation of training datasets. We explored two different data augmentation techniques, one that exploits a limited number of known unauthorized transmitters and the other that does not require any unauthorized transmitters. Our results indicate that data augmentation allows for significant increases in open set classification accuracy, especially when the authorized set is small. Another practical problem in authentication systems based on deep learning is training time when the new transmitters are added. We have developed a fast authentication algorithm based on information retrieval that uses feature vectors as RF fingerprints and locality sensitive hashing(LSH) to create a database that can quickly searched by approximate nearest neighbor algorithm. The proposed algorithm matches the accuracy of deep learning models, while allowing for more than 100x faster retraining.

Bio: Danijela Cabric is Professor in the Electrical and Computer Engineering Department at the University of California, Los Angeles. Her research interests include novel radio architectures, signal processing, communications, machine learning and networking techniques for spectrum sharing, 5G millimeter-wave, massive MIMO and IoT systems. Dr. Cabric received the Samueli Fellowship in 2008, the Okawa Foundation Research Grant in 2009, Hellman Fellowship in 2012, the National Science Foundation Faculty Early Career Development (CAREER) Award in 2012 and Qualcomm Faculty Award in 2020 and 2021. Dr. Cabric is an IEEE Fellow. Her research on deep learning based RF transmitter fingerprinting is supported by SRC/JUMP CONIX center (https://conix.io/).

About the Monthly Virtual Seminar Series:

The IEEE TCCN Security Special Interest Group conducts a monthly virtual seminar series to highlight the challenges in securing the next generation (xG) wireless networks. The talks will feature cutting edge research addressing both technical and policy issues to advance the state-of-the-art in security techniques, architectures, and algorithms for wireless communications.

Title: Brainstorming Generative Adversarial Networks (BGANs): Framework and Application to Wireless Networks

Date and Time: October 22, 2021 at 10AM EDT

Registration Process: Please register at https://tinyurl.com/2sn8tbwj

Abstract: Due to major communication, privacy, and scalability challenges stemming from the emergence of large-scale Internet of Things services, machine learning is witnessing a major departure from traditional centralized cloud architectures toward a distributed machine learning (ML) paradigm where data is dispersed and processed across multiple edge devices. A prime example of this emerging distributed ML paradigm is Google’s renowned federated learning framework. Despite the tremendous recent interest in distributed ML, remarkably, prior work in the area remains largely focused on the development of distributed ML algorithms for inference and classification tasks. In contrast, in this talk, we introduce the novel framework of brainstorming generative adversarial networks (BGANs) that constitutes one of the first implementations of distributed, multi-agent generative GAN models that does not rely on a centralized parameter server. We show how BGAN allows multiple agents to gain information from one another, in a fully distributed manner, without sharing their real datasets but by “brainstorming” their generated data samples. We then demonstrate the higher accuracy and scalability of BGAN compared to the state of the art through extensive experiments. We then illustrate how BGAN can be used to address key problems in the field of wireless communications by analyzing a millimeter wave channel modeling problem for wireless networks that rely on unmanned aerial vehicles (UAVs). We conclude this talk with an overview on the future outlook of the exciting area of distributed ML and its current and future applications in wireless systems.

Bio: Walid Saad (S’07, M’10, SM’15, F’19) received his Ph.D degree from the University of Oslo in 2010. He is currently a Professor at the Department of Electrical and Computer Engineering at Virginia Tech, where he leads the Network sciEnce, Wireless, and Security (NEWS) laboratory. His research interests include wireless networks, machine learning, game theory, security, unmanned aerial vehicles, cyber-physical systems, and network science. Dr. Saad is a Fellow of the IEEE. He is also the recipient of the NSF CAREER award in 2013 and of the Young Investigator Award from the Office of Naval Research (ONR) in 2015. He was the author/co-author of ten conference best paper awards at WiOpt in 2009, ICIMP in 2010, IEEE WCNC in 2012, IEEE PIMRC in 2015, IEEE SmartGridComm in 2015, EuCNC in 2017, IEEE GLOBECOM in 2018, IFIP NTMS in 2019, IEEE ICC in 2020, and IEEE GLOBECOM in 2020. He is the recipient of the 2015 Fred W. Ellersick Prize from the IEEE Communications Society, of the 2017 IEEE ComSoc Best Young Professional in Academia award, of the 2018 IEEE ComSoc Radio Communications Committee Early Achievement Award, and of the 2019 IEEE ComSoc Communication Theory Technical Committee. He was also a co-author of the 2019 IEEE Communications Society Young Author Best Paper and of the 2021 IEEE Communications Society Young Author Best Paper. He currently serves as an editor for several major IEEE Transactions.

About the Monthly Virtual Seminar Series:

The IEEE TCCN Security Special Interest Group conducts a monthly virtual seminar series to highlight the challenges in securing the next generation (xG) wireless networks. The talks will feature cutting edge research addressing both technical and policy issues to advance the state-of-the-art in security techniques, architectures, and algorithms for wireless communications.