Keynote Speakers

More information regarding our invited speakers will be announced in this page as it becomes available.

Noa Agmon (Bar-Ilan University)

Title: Adversarial Robotics: from Teamwork to Swarms

Developing robots for a wide range of goals requires addressing their ability to perform tasks as physical agents with specific characteristics, and the ways in which they act within and respond to their surroundings. As proximity to dangerous or hostile entities is among the foremost motives for using robots, it is therefore crucial to account for the presence of adversaries in robotic environments. The talk will describe several key research threads examining the ability of robotic teams and swarms to (strategically) handle adversity, which strongly relies on the knowledge the robots have on the environment and the opponent, and the coordination scheme between the robots.

Cynthia Dwork (Harvard University)

Title: Individual Probabilities: The Defining Problem of AI

Prediction algorithms score individuals, or individual instances, assigning to each one a number in [0,1] that is often interpreted as a probability: What are the chances that this loan will be repaid? How likely is the tumor to metastasize? What is the probability this person will commit a violent crime in the next two years? But what is the probability of a non-repeatable event? Without a satisfactory answer, we cannot even specify the goal of the ideal algorithm.

This talk will introduce Outcome Indistinguishability, a desideratum with roots in complexity theory, and situate it within the decade-old Investigation of the theory of algorithmic fairness.

Jérôme Lang (LAMSADE, CNRS, Université Paris-Dauphine, PSL)

Title: From portioning to apportioning under ordinal preferences

When a public divisible resource (such as a monetary budget) is to be divided among alternatives (such as projects), a portioning rule decides on a distribution of the budget. Examples of such portioning problems are participatory budgeting, time shares, and parliament elections. In the latter case, an integral seat assignment must be found from the (real-valued) output of the portioning rule: this is the apportionment phase. We wil show how both processes (portionment/apportionment) can be `plugged' into each other. We focus on rules for which voters have ordinal preference rankings over alternatives.


Jérôme Lang is a CNRS senior scientist. He is affiliated with the LAMSADE research institute (PSL, Université Paris-Dauphine, France). His research interests are within artificial intelligence, and more precisely computational social choice, algorithmic game theory, and knowledge representation. He was the program chair of IJCAI-ECAI-2018. He is the author or coauthor of more than 200 publications, and a co-editor of the Handbook of Computational Social Choice.

Arun Ross (Michigan State University)

Title: Modifying and Synthesizing Biometric Data

Biometrics refers to the use of physical and behavioral traits such as fingerprints, face, iris, voice and gait to recognize an individual. The biometric data (e.g., a face image) acquired from an individual may be modified for several reasons. While some modifications are intended to improve the performance of a biometric system (e.g., image enhancement), others may be intentionally adversarial (e.g., spoofing or obfuscating an identity). Furthermore, the data may be subjected to a sequence of alterations resulting in a set of near-duplicate data (e.g., applying a sequence of image filters to an input face image). In this talk, we will discuss methods for (a) detecting altered biometric data; (b) determining the relationship between near-duplicate biometric data and constructing a phylogeny tree denoting the sequence in which they were transformed; and (c) using altered biometric data to enhance privacy. The goal of the talk is to convey the dangers and, at the same time, the benefits of deliberately altered or synthesized biometric data.


Arun Ross is the John and Eva Cillag Endowed Chair in the College of Engineering and a Professor in the Department of Computer Science and Engineering at Michigan State University. He also serves as the Site Director of the NSF Center for Identification Technology Research (CITeR). He received the B.E. (Hons.) degree in Computer Science from BITS Pilani, India, and the M.S. and PhD degrees in Computer Science and Engineering from Michigan State University.

He was in the faculty of West Virginia University between 2003 and 2012 where he received the Benedum Distinguished Scholar Award for excellence in creative research and the WVU Foundation Outstanding Teaching Award.

His expertise is in the area of biometrics, computer vision and machine learning. He has advocated for the responsible use of biometrics in multiple forums including the NATO Advanced Research Workshop on Identity and Security in Switzerland in 2018. He testified as an expert panelist in an event organized by the United Nations Counter-Terrorism Committee at the UN Headquarters in 2013.

Ross serves as Associate Editor-in-Chief of the Pattern Recognition Journal, Area Editor of the Computer Vision and Image Understanding Journal and Associate Editor of IEEE Transactions on Biometrics, Behavior, and Identity Science. He has served as Associate Editor of IEEE Transactions on Information Forensics and Security, IEEE Transactions on Image Processing, IEEE Transactions on Circuits and Systems for Video Technology, ACM Computing Surveys and Image & Vision Computing Journal. He has also served as Senior Area Editor of IEEE Transactions on Image Processing.

Ross is a recipient of the NSF CAREER Award. He was designated a Kavli Fellow by the US National Academy of Sciences by virtue of his presentation at the 2006 Kavli Frontiers of Science Symposia. In recognition of his contributions to the field of pattern recognition and biometrics, he received the JK Aggarwal Prize in 2014 and the Young Biometrics Investigator Award in 2013 from the International Association of Pattern Recognition (IAPR).

Ohad Shamir (Weizmann Institute of Science)

Title: Elephant in the Room: Non-Smooth Non-Convex Optimization

It is well-known that finding global minima of non-convex optimization problems is computationally hard in general. However, the problem of finding local minima-like points (at least in terms of gradient and Hessian properties) is tractable, and received much attention in the machine learning community in recent years. The resulting literature has been largely motivated by the rising importance of non-convex optimization problems such as deep learning, but in fact, does not quite address them: Nearly all computationally efficient guarantees in this area require the objective function to be smooth, which is seldom satisfied in deep learning problems. This highlights the importance of understanding what we can do efficiently on such non-convex, non-smooth optimization problems. In this talk, I'll describe some results, challenges, and possible approaches to tackle this fundamental question.

Noam Slonim (IBM Research AI)

Title: Project Debater – an autonomous debating system

Project Debater is the first AI system that can meaningfully debate a human opponent. The system, an IBM Grand Challenge, is designed to build coherent, convincing speeches on its own, as well as provide rebuttals to the opponent's main arguments. In 2019, Project Debater competed against Harish Natarajan, who holds the world record for most debate victories, in an event held in San Francisco that was broadcasted live world-wide. In this talk I will tell the story of Project Debater, from conception to a climactic final event, describe its underlying technology and its value to business use cases, and present the results of recent systematic evaluation of the system performance.


Noam Slonim is a Distinguished Engineer at IBM Research AI. He received his doctorate from the Interdisciplinary Center for Neural Computation at the Hebrew University, under the supervision of Prof. Naftali Tishby, and held a post-doc position at the Genomics Institute at Princeton University. Noam joined the IBM Haifa Research Lab in 2007, and in 2011 he proposed to develop Project Debater. He has been serving as the Principal Investigator of the project since then.

Noga Zaslavsky (MIT)

Title: The information geometry of human pragmatic reasoning

A key aspect of language is the ability to pragmatically reason about each other’s intentions and beliefs in order to understand meaning in context. A prominent approach to pragmatics is the Rational Speech Act (RSA) framework, which formulates pragmatic reasoning as Bayesian speakers and listeners recursively reasoning about each other. While RSA enjoys broad empirical support, its predictions have been explored mainly numerically rather than analytically, and on that basis it has been conjectured that the RSA recursion locally increases communicative utility. In this talk, I will present an information-geometric analysis of the RSA framework that sheds new light on the principles that may govern pragmatic reasoning. First, I will show that RSA’s recursive reasoning implements an alternating minimization algorithm. Rather than optimizing utility, however, it optimizes a tradeoff between utility and communicative effort. Second, I will show that RSA can be grounded in Rate–Distortion theory, yielding RD-RSA: a principled model of pragmatic reasoning that has similar predictive power for human behavior as RSA while avoiding a provable bias of RSA toward random utterance production. These results suggest that pressure for efficient compression may give rise to pragmatic reasoning, leading to a new approach for endowing artificial agents with human-like pragmatic skills without direct human supervision.


Judea Pearl 85th Birthday Celebration

Interview of Judea Pearl by Stephen Wolfram. The interview is now available on YouTube.