Speakers

Associate Professor, University of Texas Health Science Center at Houston (UTHealth ), USA

Title: Honor Ethics: The challenge of globalizing value alignment in AI

Abstract: Ethical failures in AI have instigated a push for fair, responsible, and trustworthy AI. However, we contend that much of the ethical framing for AI comes from a limited WEIRD (Western, Educated, Industrialized, Rich, Democratic) perspective. In this talk, we put on a cultural anthropologist’s hat, arguing that the concept of honor would expand the discussion to non-WEIRD global communities. We first describe honor according to recent empirical and philosophical scholarship. Then, we review “consensus” principles for AI ethics framed from an honor-based perspective. A better appreciation of these marginalized conceptions of honor could, we hope, lead to AI systems that better reflect the needs and values of users around the globe.

Bio: Dr. Stephen Wu is an Associate Professor at the University of Texas Health Science Center at Houston (UTHealth) and the Head of Research at Tibyan (تبيرن, a new NGO focused on “Clarifying AI”). After his PhD in natural language processing at the University of Minnesota, he worked in medical informatics at Mayo Clinic and Oregon Health & Science University. He and his family have been living in North Africa since 2019. Current research projects include the ethics of AI, electronic medical records, and lower-resourced languages.

PhD Student at New York University, and Visiting Researcher at Meta AI

Title: Are the Marginal Likelihood and PAC-Bayes Bounds the right proxies for Generalization?

Abstract: How do we compare between hypotheses that are entirely consistent with observations? The marginal likelihood, which represents the probability of generating our observations from a prior, provides a distinctive approach to this foundational question. We first highlight the conceptual and practical issues in using the marginal likelihood as a proxy for generalization. Namely, we show how the marginal likelihood can be negatively correlated with generalization and can lead to both underfitting and overfitting in hyperparameter learning. We provide a partial remedy through a conditional marginal likelihood, which we show to be more aligned with generalization, and practically valuable for large-scale hyperparameter learning, such as in deep kernel learning. PAC-Bayes bounds are another expression of Occam’s razor where simpler descriptions of the data generalize better. While there has been progress in developing tighter PAC-Bayes bounds for deep neural networks, these bounds tend to be uninformative about why deep learning works. In this talk, I will also present our compression approach based on uantizing neural network parameters in a linear subspace, which profoundly improves on previous results to provide state-of-the-art generalization bounds on a variety of tasks. We use these tight bounds to better understand the role of model size, equivariance, and the implicit biases of optimization for generalization in deep learning. Notably, our work shows that large models can be compressed to a much greater extent than previously known and argues for data-independent bounds in explaining generalization. Finally, I will discuss the connection between the marginal likelihood and PAC-Bayes bounds for model selection.

Bio: Sanae Lotfi is a PhD student at NYU advised by Professor Andrew Gordon Wilson, and a visiting researcher at Meta AI where she works with Brandon Amos. She is currently interested in understanding and quantifying the generalization properties of deep learning models. More broadly, her research interests include robustness to distribution shift, Bayesian learning, large-scale optimization, and time series forecasting. Sanae's PhD research has been recognized with an ICML Outstanding Paper Award and is generously supported by the Microsoft Research PhD Fellowship, the DeepMind Fellowship, the Meta AI Mentorship Program and the NYU Center for Data Science Fellowship. Prior to joining NYU, Sanae obtained a Master’s degree in applied mathematics from Polytechnique Montreal, where she worked on designing stochastic first and second order algorithms with compelling theoretical and empirical properties for machine learning and large-scale optimization. Sanae also holds a Master's degree in general engineering and applied mathematics from CentraleSupélec.

Lead Researcher, Technology Innovation Institute, United Arab Emirates

Title: Blockchain and AI for Cyber Security in the Era of IoT applications: Research challenges

Abstract: The 5th revolution of the industrial era – or Industry 5.0 is the new industry trend that defines the Smart Factory concept. This concept is based on emerging technologies such as 5G/6G communications, Fog computing, Drones, Cloud computing, Blockchain, Artificial Intelligence, Deep learning, etc. To provide the optimization of operations and the reduction of costs, these technologies are employed to establish a connection between machines and the Internet, through the Internet of Things, and to collect information in the Cloud and Edge and then process them using artificial intelligence algorithms. However, with thousands of IoT-based devices deployed in the open field, there are many new cyber security threats in Industry 5.0. When an adversary attempts to penetrate the network, it uses several different approaches such as DDoS attacks, scanning attacks, false data injection attacks, to disrupt the functioning of the IoT-based devices. To protect Industry 5.0 from destruction, change, unauthorized access, or attack, Security researchers propose the use of an intrusion detection system (IDS) combined the blockchain technology. The IDS system is a mechanism monitoring the network traffic, which is used to detect suspicious or abnormal activities and then enables preventive measures on the intrusion risks. The blockchain technology is used to detect fraudulent transactions. In this talk, I will first discuss blockchain technology as well as artificial intelligence for cyber security in IoT applications. Next, I will move to survey some solutions developed by my research group, intended to provide security to IoT applications. Finally, I will conclude by highlighting topic-related research directions.

Bio: Dr. Mohamed Amine Ferrag received the Bachelor’s, Master’s, Ph.D., and Habilitation degrees in computer science from Badji Mokhtar—Annaba University, Annaba, Algeria, in June, 2008, June, 2010, June, 2014, and April, 2019, respectively. From 2014 to 2022, he was an Associate Professor with the Department of Computer Science, Guelma University, Algeria. From 2019 to 2022, he was a Visiting Senior Researcher with the NAU-Lincoln Joint Research Center of Intelligent Engineering, Nanjing Agricultural University, China. Since 2022, he has been the Led Researcher with Artificial Intelligence & Digital Science Research Center, Technology Innovation Institute, Abu Dhabi, United Arab Emirates. His research interests include wireless network security, network coding security, applied cryptography, blockchain technology, and AI for cyber security. He has published over 100 papers in international journals and conferences in the above areas. He has been conducting several research projects with international collaborations on these topics. He was a recipient of the 2021 IEEE TEM Best Paper Award. He is featured in Stanford University’s list of the world’s Top 2% scientists for the years 2020, 2021, and 2022. He is a Senior Member of the Institute of Electrical & Electronic Engineers (IEEE) and a member of the Association for Computing Machinery (ACM).