CROSSING Conference 2023

Schedule

Time Speakers / Talks
08:30 – 09:30 Registration and welcome
09:30 – 09:45 Opening
Prof. Ahmad-Reza Sadeghi (TU Darmstadt)
09:45 – 10:30 Future trends and applications of large scale models: A view from the trenches
Dr. Saurabh Tiwary (Microsoft Turing)
10:30 – 11:05 Automated Cryptographically-Secure Private Computing: From Logic and Mixed-Protocol Optimization to Centralized and Federated ML Customization
Prof. Farinaz Koushanfar (UCSD)
11:05 – 11:30 Coffee Break
11:30 – 12:05 Enhanced NFTs
Prof. Ivan Visconti (Salerno University)
12:05 – 12:40 Covert & Side Stories: Threats Evolution in Traditional and Modern Technologies
Prof. Mauro Conti (University of Padua)
12:40 – 14:10 Lunch
14:10 – 14:45 Side-channel analysis of cryptographic implementations: Lessons learned and future directions
Prof. Lejla Batina (Radbaud University)
14:45 – 15:20 AI and Cybersecurity: A Perfect Couple…or not?
Prof. Stjepan Picek (Radbaud University)
15:20 – 15:55 Network-level Adversaries in Federated Learning
Prof. Cristina Nita-Rotaru (Northeastern University)
15:55 – 16:15 Coffee Break
16:15 – 16:50 Trustworthy Cyber-Physical Critical Infrastructures via Physics-Aware and AI-Powered Security
Prof. Saman Zounouz (Georgia Institute of Technology)
16:50 – 17:25 Exploring Two frontier problems in hardware security using AI
Prof. Jeyavijayan Rajendran (Texas A&M University)
from 19:00 Conference Dinner
(included in the conference fee)
Time Speakers / Talks
09:30 – 09:45 Opening
Prof. Ahmad-Reza Sadeghi (TU Darmstadt)
09:45 – 10:30 Business Risk to Application Security
Lucas von Stockhausen (Synopsys SIG)
10:30 – 11:05 Quantum Computation and Simulation with Laser-Cooled Neutral Atoms
Prof. Gerhard Birkl (TU Darmstadt)
11:05 – 11:30 Coffee break
11:30 – 12:05 The Evolving Landscape of Datatypes in AI
Dr. Bita Rouhani (Microsoft Azure)
12:05 – 12:30 AI – from lab to market between strangulation and freedom
Dr. h.c. Thomas Sattelberger
12:30 – 14:10 Lunch
14:10 – 14:45 A dive into confidential computing in the distributive computing paradigm
Dr. Yin Tan (Huawei)
14:45 – 15:20 Securing location-based mobile computing
Prof. Panos Papadimitratos (KTH Royal Institute of Technololgy)
15:20 – 15:55 Confronting Adaptive Attackers in Federated Learning: Challenges and Countermeasures
Prof. Alexandra Dmitrienko (University of Würzburg)
15:55 – 16:15 Coffee break
16:15 – 17:20 Panel Discussion
Panelists:
  • Prof. Farinaz Koushanfar (University of California San Diego)
  • Dr. Durga Ramachandran (Riscure)
  • Lucas von Stockhausen (Synopsys SIG)
  • Dr. Yin Tan (Huawei)
17:20 – 17:55 Critical Infrastructure Security
Manuel Atug (AG KRITIS)

Abstracts, Slides and Speaker Bios

Talk: Critical Infrastructure Security
download slides (opens in new tab)


Abstract
The speaker will discuss about the following topics and questions: What is critical infrastructure (KRITIS)? How do critical infrastructures look like in the Cybersecurity point of view of their IT and OT world? What kind of Cyber-physical attacks and ICS attacks were reelvant in the past and why? How do solutions look like to secure critical infrastructure? Is the legal setup of e.g. the IT Security Act and the upcomming EU NIS2 sufficient? What is sustainable digitization and how can we do that?

Speaker Bio
Manuel 'HonkHase' Atug has been working in information security and critical infrastructure protection as a consultant and auditor for well over 23 years, covering the topics of KRITIS, hackback, ethics, hybrid warfare, cyber resilience, population and disaster management. HonkHase holds a degree in computer science, a Master of Science in Applied IT Security and is an engineer. Furthermore, he is founder and spokesperson of the independent working group Critical Infrastructures (AG KRITIS) and co-founder of the independent working group Sustainable Digitalization (AGND) together with Caroline Krohn. He is active on the web as @HonkHase.

Side-channel analysis of cryptographic implementations: Lessons learned and future directions

Abstract
Side-channel analysis has changed the field of cryptography and it became the most common cause of real-world security applications failing today. It has also shaped the way crypto competitions are run such as recently finished NIST Post-quantum and Lightweight crypto standardization processes. In this talk we give an overview of side-channel attacks on implementations of cryptography and countermeasures. We discuss the ways in which Machine learning and AI changed the side-channel analysis landscape and attackers’ capabilities in particular. We survey several examples of AI assisting with leakage evaluation and discuss the impact of it on the field and security evaluations in particular. Finally, we also describe the way side-channel analysis threatens AI implementations e.g. neural nets architectures that are commonly used in practice. In the end, we identify some avenues for future research.

Speaker Bio
Lejla Batina is a professor in embedded systems security at the Radboud University in Nijmegen, the Netherlands. She received her Ph.D. from KU Leuven, Belgium (2005) and before that she worked as a cryptographer for SafeNet B.V. in The Netherlands (2001–2003).

She has coauthored more than 150 refereed articles and her current research interests include physical attacks on cryptographic implementations and the impact of AI on hardware security.

She is a senior member of IEEE and an Editorial board member of top journals in security, such as IEEE Transactions on Information Forensics and Security and ACM Transactions on Embedded Computing Systems. She was program co-chair of CHES 2014, ACM WiSec 2021, Africacrypt 2022, SPACE 2020-2022 and she co-organized (as general chair) IACR flagship conferences like EUROCRYPT and Real-world crypto symposium (RWC). Her research group at Radboud consists of 10+ researchers and 12 Ph.D. students have so far graduated under her supervision.

Quantum Computation and Simulation with Laser-Cooled Neutral Atoms

Abstract
Through significant recent progress, regular registers of individual neutral atoms have established themselves as a third auspicious techological platform for quantum information processing complementing superconducting circuits and ion traps. Hundreds of quantum bits (qubits) are available routinely. Work towards thousands of qubits is in progess and 100 thousands are within reach by extenion of currently available technology. One-qubit gate operations are induced by electromagnetic fields and two-qubit gates are achieved by direct mutual interaction of neighboring atoms in highly-energetic internal states.
I will discuss the physical principles and technological basis for quantum information processing with neutral atoms cooled close to absolute zero temperature, held in vacuum, and trapped by light. I will present the DArmstadt-Neutral-Atom-Quantum-Technology-Platform (DANAQTP) as a scalable quantum information processing architecture based on advanced optical and quantum optical technologies. Recent progress and future perspectives will complement the presentation.

Speaker Bio
Gerhard Birkl is a professor at the Institute for Applied Physics (IAP) at TU Darmstadt. He leads the research group “Atoms-Photons-Quanta” and conducts experimental and theoretical research in the fields of Quantum Optics, Quantum Information, Quantum Technology and Atomic Physics.
After receiveing his doctorate from Ludwig-Maximilians-Universität München, Gerhard Birkl worked as a researcher at Max-Planck-Institut für Quantenoptik and was a postdoc with Nobel laureats Prof. William D. Phillips (National Institute of Science and Technology, Maryland, USA) and Prof. Alain Aspect (Institute d’Optique Théoretique et Appliquée, Orsay, France). Before joining TU Darmstadt, he held a private lecturer position at Leibniz Universität Hannover.

Talk: Covert & Side Stories: Threats Evolution in Traditional and Modern Technologies
download slides (opens in new tab)


Abstract
Alongside traditional Information and Communication Technologies, more recent ones like Smartphones and IoT devices also became pervasive. Furthermore, all technologies manage an increasing amount of confidential data. The concern of protecting these data is not only related to an adversary gaining physical or remote control of a victim device through traditional attacks, but also to what extent an adversary without the above capabilities can infer or steal information through side and covert channels!

In this talk, we survey a corpus of representative research results published in the domain of side and covert channels, ranging from TIFS 2016 to more recent Usenix Security 2022, and including several demonstrations at Black Hat Hacking Conferences. We discuss threats coming from contextual information and to which extent it is feasible to infer very specific information. In particular, we discuss attacks like inferring actions that a user is doing on mobile apps, by eavesdropping their encrypted network traffic, identifying the presence of a specific user within a network through analysis of energy consumption, or inferring information (also key one like passwords and PINs) through timing, acoustic, or video information.

Speaker Bio
Mauro Conti is Full Professor at the University of Padua, Italy. He is also affiliated with TU Delft and University of Washington, Seattle. He obtained his Ph.D. from Sapienza University of Rome, Italy, in 2009. After his Ph.D., he was a Post-Doc Researcher at Vrije Universiteit Amsterdam, The Netherlands. In 2011 he joined as Assistant Professor at the University of Padua, where he became Associate Professor in 2015, and Full Professor in 2018. He has been Visiting Researcher at GMU, UCLA, UCI, TU Darmstadt, UF, and FIU. He has been awarded with a Marie Curie Fellowship (2012) by the European Commission, and with a Fellowship by the German DAAD (2013). His research is also funded by companies, including Cisco, Intel, and Huawei.

His main research interest is in the area of Security and Privacy. In this area, he published more than 500 papers in topmost international peer-reviewed journals and conferences. He is Editor-in-Chief for IEEE Transactions on Information Forensics and Security, Area Editor-in-Chief for IEEE Communications Surveys & Tutorials, and has been Associate Editor for several journals, including IEEE Communications Surveys & Tutorials, IEEE Transactions on Dependable and Secure Computing, IEEE Transactions on Information Forensics and Security, and IEEE Transactions on Network and Service Management. He was Program Chair for TRUST 2015, ICISS 2016, WiSec 2017, ACNS 2020, CANS 2021, WiMob 2023 and ESORICS 2023, and General Chair for SecureComm 2012, SACMAT 2013, NSS 2021 and ACNS 2022. He is Fellow of the IEEE, Fellow of the AAIA, Senior Member of the ACM, and Fellow of the Young Academy of Europe.

Confronting Adaptive Attackers in Federated Learning: Challenges and Countermeasures
download slides (opens in new tab)

Abstract
Federated Learning (FL) represents a decentralized strategy in Machine Learning (ML) that greatly enhances the privacy of the training process. Within FL, a variety of clients autonomously refine their models using their distinct datasets over numerous training rounds. After each training round, locally trained models are transmitted to a central server for aggregation, which is trusted to aggregate all the models into a global model, yet has no access to the underlying training data as they remain securely stored on the clients' end. Recent research has shown, however, that despite being more-privacy-preserving than traditional ML, FL is similarly prone to poisoning attacks, which can reduce performance of the aggregated global model, or add adversarial behavior, i.e., a backdoor, that can be triggered with appropriately crafted inputs.

In this talk, we will delve into the persistent challenges of training effective global models, while detecting poisoned local model updates and removing them from aggregation. Particular issue here is the ability to distinguish between poisoned models and benign, but uncommon models, e.g., trained on datasets featuring distinct data distributions. Another ongoing challenge is posed by adaptive attackers, who, once familiar with detection methods, can strategically introduce an additional training loss to minimize any shifts in detection metrics, thereby successfully avoiding identification.We will present our two very recent defense methods, CrowdGuard and MESAS, which advance the state-of-the-art by providing enhanced resilience against adaptive attackers and ability to function in scenarios with variously distributed datasets. We will also highlight potentials for improvements and outline promising directions to foster productive discussions in the research community.

Speaker Bio
Alexandra Dmitrienko is an Associate Professor and head of the Secure Software Systems group at the University of Wuerzburg in Germany. Before taking her current faculty position in 2018, she collected an extensive background in security institutions in Germany and Switzerland, including Ruhr-University Bochum (2008-2011), Fraunhofer Institute for Information Security in Darmstadt (2011-2015), and ETH Zurich (2016-2017). She earned her PhD in Security and Information Technology from TU Darmstadt (2015), where her dissertation focused on the security and privacy of mobile systems and applications, and was recognized with awards from the European Research Consortium in Informatics and Mathematics (ERCIM STM WG 2016 Award) and Intel (Intel Doctoral Student Honor Award, 2013). Over the years, her research interests spawned across various topics such as secure software engineering, systems security and privacy, security and privacy of mobile, cyber-physical, and distributed systems. Today, her recent research also largely focuses on security and privacy aspects of Artificial Intelligence methods.

Automated Cryptographically-Secure Private Computing: From Logic and Mixed-Protocol Optimization to Centralized and Federated ML Customization

Abstract

Over the last four decades, much research effort has been dedicated to designing cryptographically-secure methods for computing on encrypted data. However, despite the great progress in research, adoption of the sophisticated crypto methodologies has been rather slow and limited in practical settings. Presently used heuristic and trusted third party solutions fall short in guaranteeing the privacy requirements for the contemporary massive datasets, complex AI algorithms, and the emerging collaborative/distributed computing scenarios such as blockchains.

In this talk, we outline the challenges in the state-of-the-art protocols for computing on encrypted data with an emphasis on the emerging centralized, federated, and distributed learning scenarios. We discuss how in recent years, giant strides have been made in this field by leveraging optimization and design automation methods including logic synthesis, protocol selection, and automated co-design/co-optimization of cryptographic protocols, learning algorithm, software, and hardware. Proof of concept would be demonstrated in the design of the present state-of-the-art frameworks for cryptographically-secure deep learning on encrypted data. We conclude by discussing the practical challenges in the emerging private robust learning and distributed/ federated computing scenarios as well as the opportunities ahead.

Speaker Bio
Farinaz Koushanfar is the Henry Booker Scholar Professor of ECE at the University of California San Diego (UCSD), where she is also the founding co-director of the UCSD Center for Machine-Intelligence, Computing & Security (MICS). Her research addresses several aspects of secure and efficient computing, with a focus on hardware and system security, robust machine learning under resource constraints, intellectual property (IP) protection, as well as practical privacy-preserving computing. Dr. Koushanfar is a fellow of the Kavli Frontiers of the National Academy of Sciences and a fellow of IEEE/ACM. She has received a number of awards and honors including the Presidential Early Career Award for Scientists and Engineers (PECASE) from President Obama, the ACM SIGDA Outstanding New Faculty Award, Cisco IoT Security Grand Challenge Award, MIT Technology Review TR-35, Qualcomm Innovation Awards, Intel Collaborative Awards, Young Faculty/CAREER Awards from NSF, DARPA, ONR and ARO, as well as several best paper awards.

Talk: Network-level Adversaries in Federated Learning
download slides (opens in new tab)


Abstract
Federated learning is a popular strategy for training models on distributed, sensitive data, while preserving data privacy.
In both centralized and peer-to-peer architectures communication between participants (clients and server or peers) plays a critical role for the learning task performance. We highlight how communication introduces another vulnerability surface in federated learning and study the impact of network-level adversaries on training federated learning models. We first focus on centralized architectures and show that attackers dropping the network traffic from carefully selected clients can significantly decrease model accuracy on a target population. We then show attacks in the context of peer-to-peer architectures. We conclude by showing the effectiveness of our server-side defense which mitigates the impact of our attacks by identifying and up-sampling clients likely to positively contribute towards target accuracy.

Speaker Bio
Cristina Nita-Rotaru is a Professor of Computer Science in the Khoury College of Computer Sciences at Northeastern University where she leads the Network and Distributed Systems Security Laboratory (NDS2). Prior to joining Northeastern she was a faculty in the Department of Computer Science at Purdue University (2003 – 2015). She served as Associate Dean of Faculty at Northeastern University (2017 – 2020) and as an Assistant Director for CERIAS at Purdue University (2011 – 2013). Her research lies at the intersection of security, distributed systems, and computer networks. The overarching goal of her work is designing and building secure and resilient distributed systems and network protocols, with assurance that the deployed implementations provide their security, resilience, and performance goals. Her work received several best paper awards in NETYS 2023, ACM SACMAT 2022, IEEE SafeThings 2019, NDSS 2018, ISSRE 2017, DSN 2015, two IETF/IRTF Applied Networking Research Prize in 2018 and 2016, and Test-of-Time award in ACM SACMAT 2022. She is a recipient of the NSF Career Award in 2006.

Securing location-based mobile computing

Abstract
A broad gamut of Internet of Things (IoT) and mobile applications are location-based: their operation relies on real-time, precise position information, or they collect location-specific information on the user (physical) environment. They have gained wide popularity, offering valuable services to users and systems. This brings forth a dual challenge: how to secure position information and how to safeguard the system from misbehaving data-collecting devices/users. In this talk, we discuss briefly the two problems and relevant solutions. First, methods to secure Global Navigation Satellite System (GNSS)-based position (and time) services; second, vulnerabilities of user-contributing location based applications and an architecture and protocols to thwart such attacks.

Speaker Bio
Panos Papadimitratos is a Professor at KTH, Stockholm, Sweden, leading the Networked Systems Security (NSS) Group. He earned his PhD from Cornell University, Ithaca, NY, USA. After holding positions at Virginia Tech, EPFL, and a visiting post at PoliTo, he joined KTH and established the NSS Group. He is a Fellow of the Young Academy of Europe (YAE) and an IEEE Fellow, Class of 2020.

Talk:AI and Cybersecurity: A Perfect Couple…or not?
download slides (opens in new tab)


Abstract
Today, AI represents a go-to approach for various cybersecurity applications like side-channel attacks, attacks on physically unclonable functions, fuzzing, malware analysis, etc. Moreover, recent years witnessed a considerable increase in the number of works examining the security and privacy of AI. As such, it is evident that AI and cybersecurity have become a topic of utmost importance. Unfortunately, despite the results in the last few years, many unanswered questions remain. In this talk, we will briefly discuss several success stories, and afterward, we will examine the challenges to be addressed.

Speaker Bio
Stjepan Picek is an associate professor at Radboud University, The Netherlands. His research interests are security/cryptography, machine learning, and evolutionary computation. Prior to the associate professor position, Stjepan was an assistant professor at TU Delft, and a postdoctoral researcher at MIT, USA and KU Leuven, Belgium. Stjepan finished his PhD in 2015 with a topic on cryptology and evolutionary computation techniques. Stjepan also has several years of experience working in industry and government. Up to now, Stjepan has given more than 30 invited talks and published more than 150 refereed papers. He is a program committee member and reviewer for a number of conferences and journals, and a member of several professional societies. His work has been featured in the mainstream media and on popular technology blogs.

Exploring Two frontier problems in hardware security using AI

Abstract
Hardware is at the heart of computing systems. For decades, software was considered error-prone and vulnerable. However, recent years have seen a rise in attacks exploiting hardware vulnerabilities and exploits. Such vulnerabilities are prevalent in hardware for several reasons: First, the existing functional verification and validation approaches do not account for security, motivating the need for new and radical approaches such as hardware fuzzing. Second, existing defense solutions, mostly based on heuristics, do not undergo rigorous red-teaming exercises like cryptographic algorithms; I will talk about how emerging artificial intelligence (AI) can rapidly help red-team such techniques. Last and most important, students and practitioners who are typically trained in designing, testing, and verification are not rigorously trained in cybersecurity -- for many reasons, including a lack of resources, time, and methodologies; I will talk about how AI can be incorporated into (hardware) cybersecurity education.

Speaker Bio
Jeyavijayan (JV) Rajendran is an ASCEND Fellow and an Assistant Professor in the Department of Electrical and Computer Engineering at the Texas A&M University. He obtained his Ph.D. degree from New York University in August 2015. His research interests include hardware security and computer security. His research has won the NSF CAREER Award in 2017, ONR Young Investigator Award in 2022, the IEEE CEDA Ernest Kuh Early Career Award in 2021, the ACM SIGDA Outstanding Young Faculty Award in 2019, the Intel Academic Leadership Award, the ACM SIGDA Outstanding Ph.D. Dissertation Award in 2017, and the Alexander Hessel Award for the Best Ph.D. Dissertation in the Electrical and Computer Engineering Department at NYU in 2016, along with several best student paper awards. He organizes and has co‐founded Hack@DAC, a student security competition co-located with DAC, and SUSHI.

The Evolving Landscape of Datatypes in AI

Abstract
The rapid emergence of disruptive AI capabilities is predominantly fueled by the principle of scale. While this scaling brings about remarkable advancements, it also places considerable demands on computation and energy resources. To address these challenges, researchers in the industry and academia have explored various techniques, with quantization of tensor values standing out as a vital approach. In this talk, we introduce an end-to-end framework tailored for comparing and evaluating a diverse array of narrow-precision formats within the realm of deep learning. We further discuss the effectiveness of a new data type, known as microexponent (MX) formats, which not only reduces user friction to leverage quantization but, for the first time, enables sub-8-bit training and inference without necessitating calibration or alterations to training recipes. MX, featuring multi-level hardware-assisted fine-grain scaling, consistently delivers strong performance across a spectrum of real-world models, including large-scale generative pretraining and recommendation systems. These findings, in turn, unlock a more sustainable solution for AI at scale while ensuring broader accessibility and utilization as the technology evolves over time.

Speaker Bio
Bita Rouhani is a Principal R&D manager at Microsoft, working at the intersection of AI algorithms and architecture. She leads a highly interdisciplinary team and serves as the AI data science lead for Azure AI Supercomputer. Her work in advanced AI datatypes and algorithms has been deployed at large scales across multiple generations of AI hardware.

Business Risk to Application Security

Abstract
Thomas Sattelberger will discuss current AI-Regulation approaches in Europe and in the US and will take a stand for “light touch” regulation. He also will comment on the disappointing AI – situation in Germany and Europe: in politics, SME’s , research transfer and spin outs out of academia. Thomas is advocating for more freedom, no matter if regulatory sandboxes, worry-free IP transfer or university enterprise zones

Speaker Bio
Dr. h.c. Thomas Sattelberger was a member of the German Bundestag from October 2017 to August 2022 and Parliamentary State Secretary to the Federal Minister of Education and Research from 2021 to 2022. Prior to that, he was a member of the executive boards of German DAX companies for many years, including as Chief Human Resources Officer at Deutsche Telekom and Continental AG. Among other things, he founded the national initiative “MINT Zukunft schaffen” (Creating STEM Future) and was its chairman for many years.

Talk: Business Risk to Application Security
download slides (opens in new tab)


Abstract
A risk management approach can help organizations to shape their AppSec program to protect business and maintain the trust of your users, even as the pace, complexity, and security risks of software delivery increases.

Speaker Bio
Lucas has over 15 years of experience in Application Security with a deep knowledge of Static, Dynamic and Interactive Application security testing as well as Software Composition Analysis and Runtime Application Self Protection technologies.

Working as Sales Engineer, Solution Architect, Solution Advisor, Application Security Strategist and Product Manager for Application Security Products Lucas experienced Application Security in all different stages. He saw all the issues a customer can experience during the journey of Application Security as well as the benefits which a successful implementation of an Application Security Testing framework can bring. This becomes especially true if customers are on their journey to implement Security into DevOps which is commonly called DevSecOps.

After joining Synopsys Lucas successfully build a Solution Center of Excellence to work with customers on a successful implementation of their Application Security initiative. He is now leading the International Sales Engineering group across EMEA, Japan, China and all other areas in Asia and Australia.

A dive into confidential computing in the distributive computing paradigm

Abstract
To meet the comprehensive requirements of data protection, confidential computing has been studied and applied extensively. Recently, the accelerator hardware based confidential computing received significant attentions due to the importance of data and model protection for AI. In this talk, we will introduce our research of confidential computing in the distributive computing paradigm and share our future plans.

Speaker Bio
Dr. Yin Tan received his Ph.D. from Peking University. He worked with National University of Singapore as a postdoc, and then with University of Waterloo as the assistant professor. His major research areas are cryptography, operating system security, trusted execution environment and trusted computing. Currently he is leading the global research of system security for Huawei.

Future trends and applications of large scale models: A view from the trenches

Abstract
Large language models (LLMs) are powerful tools for natural language processing that have rapidly gained attention in the past year. Yet, the foundations of deep learning have been steadily evolving for much longer, leading to remarkable advances and real-world impacts. In this talk, we will explore how some of these deep learning techniques are applied to create the Microsoft CoPilot family, a suite of intelligent assistants for various domains and tasks. We will also reflect on some of the historical developments and future directions of LLMs and deep learning in general. Finally, we will highlight some of the current challenges and open research questions that invite further innovation and exploration.

Speaker Bio
Saurabh Tiwary is Corporate Vice President & Technical Fellow at Microsoft. He leads Microsoft Turing which is a deep learning initiative at Microsoft that he founded in 2015 from scratch (literally setting up the first significant batch of GPUs) to building some of the largest language models in the world. Deep learning models from his team are today serving 100s of millions of users. His current focus is on scaling deep learning efforts at Microsoft to improve various products (Bing, M365, Outlook, Word, Teams, PowerPoint, etc.) by helping drive development of

  • large scale training,
  • state of the art models for NLP and computer vision, and
  • efficient inferencing across different hardware choices — CPUs, GPUs, FPGAs and custom ASICs.

He has previously worked at Cadence Research Labs and Google. At Google, he worked on pagerank, synonyms, and intent detection for web pages as part of the search quality team. He also incubated a project which graduated as Google Cloud Talent Solution. He has a Bachelors degree from IIT Kanpur. He obtained his Masters and PhD from Carnegie Mellon University and an MBA from Haas School of Business at UC Berkeley.

Talk: Enhanced NFTs
download slides (opens in new tab)


Abstract
NFTs are among the most relevant applications of blockchain technology. They have generated great enthusiasm due to their flexibility, but also some serious criticism due to various limitations of current instantiations.In this talk we will present enhanced notions and constructions of NFTs that bypass some of the above limitations.

Speaker Bio
Ivan Visconti is a professor of Computer Science at University of Salerno. His research focuses on cryptographic protocols and blockchain technology. He has served multiple times in program committees of prestigious IACR conferences. His research from 2018 to 2021 has been funded by the EU-H2020 project “Privacy-Enhancing Cryptography in Distributed Ledgers”.

Trustworthy Cyber-Physical Critical Infrastructures via Physics-Aware and AI-Powered Security

Abstract

Critical cyber-physical infrastructures, such as the power grid and manufacturing, integrate networks of computational and physical processes to provide people across the globe with essential functionalities and services. Protecting these critical infrastructures’ security against adversarial parties is a vital necessity because the failure of these systems would have a debilitating impact on economic security, public health, and safety. Our research aims at the provision of real-world solutions to facilitate the secure and reliable operation of next-generation critical infrastructures. This requires interdisciplinary research efforts across adaptive systems and network security, cyber-physical systems, and trustworthy real-time detection and response mechanisms.

In this talk, I will focus on real past and potential future threats against critical infrastructures and embedded controllers, and discuss the challenges in the design, implementation, and analysis of security solutions to protect cyber-physical platforms. I will introduce novel classes of working systems that we have developed to overcome these challenges. In particular, I will present our solutions for security verification of cyber-physical controllers for safe power grid and avionics operations. I will review our results to protect additive manufacturing and 3D printer security to ensure the structural integrity of the ultimate printed objects. Finally, I will briefly talk about our recent efforts in security monitoring of the controller side-channel signals for online attack detection purposes.

Speaker Bio
Saman Zonouz is an Associate Professor at Georgia Tech in the Schools of Cybersecurity and Privacy (SCP) and Electrical and Computer Engineering (ECE). Saman directs the Cyber-Physical Security Laboratory (CPSec). His research focuses on security and privacy research problems in cyber-physical systems including attack detection and response capabilities using techniques from systems security, control theory and artificial intelligence. His research has been awarded by Presidential Early Career Awards for Scientists and Engineers (PECASE) by the United States President, the NSF CAREER Award in Cyber-Physical Systems (CPS), Significant Research in Cyber Security by the National Security Agency (NSA), and Faculty Fellowship Award by the Air Force Office of Scientific Research (AFOSR). His research group has disclosed several security vulnerabilities with published CVEs in widely-used industrial controllers such as Siemens, Allen Bradley, and Wago.

Saman is currently a Co-PI on President Biden’s American Rescue Plan $65M Georgia AI Manufacturing (GA-AIM) project. Saman was invited to co-chair the NSF CPS PI Meeting as well as the NSF CPS Next Big Challenges Workshop. Saman has served as the chair and/or program committee member for several conferences (e.g., IEEE S&P, USENIX, CCS, NDSS, DSN, and ICCPS). Saman obtained his Ph.D. in Computer Science from the University of Illinois at Urbana-Champaign.