Peter Shor is the 2025 Shannon Lecturer.
The ISIT 2025 plenary speakers are (in alphabetical order):
- Nilanjana Datta, University of Cambridge
- Alexander Rakhlin, MIT
- Éva Tardos, Cornell University
- En-Hui Yang, University of Waterloo
Shannon Lecture
Peter Shor (MIT, USA)
Date: Thursday (6/26)
Venue: Lydia Mendelssohn Theatre
Title: Quantum Error Correcting Codes And Quantum Channel Capacities
Abstract: Often, when quantum mechanics is introduced into a theory, the theory changes radically. This happens with quantum information theory. We will discuss quantum error correcting codes and quantum channel capacities. While there is essentially only one classical capacity, there are many different quantum capacities, and they do not behave as nicely as classical capacity. Quantum error correcting codes must satisfy more constraints than classical ones, which means they are quite a bit harder to construct. We still don't have efficient quantum error correcting codes that asymptotically approach the quantum capacity. We will discuss some of the parallels and the differences between classical and quantum information theory.
Bio: Peter Shor is Morss Professor of Applied Mathematics since 2003. He received the B.A. in mathematics from Caltech in 1981, and the Ph.D. in applied mathematics from MIT in 1985, under the direction of Tom Leighton. Following a postdoctoral fellowship at MSRI, he joined AT\&T. He was a member of its Research staff, 1986-2003. He joined the MIT faculty in applied mathematics as full professor in 2003. Professor Shor's research interests are in theoretical computer science: currently on algorithms, quantum computing, computational geometry and combinatorics. In 1998, Peter Shor received the Nevanlinna Prize and the International Quantum Communication Award. He also received the Dickson Prize in Science from Carnegie-Mellon in 1998. He was awarded the Gödel Prize of the ACM and a MacArthur Foundation Fellowship in 1999. He received the King Faisal International Prize in Science in 2002, and was named one of Caltech's Distinguished Alumni in 2007. He is a member of the National Academy of Science (2002), and fellow of the American Academy of Arts and Sciences (2011). In 2017, Professor Shor received the Dirac Medal of the International Centre for Theoretical Physics. He also received the 2017 IEEE Information Theory Society Paper Award, jointly with Charles Bennett, Igor Devetak, Aram Harrow, and Andreas Winter for the paper "The Quantum Reverse Shannon Theorem and Resource Tradeoffs for Simulating Quantum Channels" which appeared in the IEEE Transactions on Information Theory, vol. 60, no. 5, pp. 2926–2959, May 2014. In 2018, Shor received the IEEE Eric E. Sumner Award, for Outstanding Contributions to Communications Technology. He also received the 2018 Micius Quantum Prize in April 2019. In May 2022, Shor was named the recipient of MIT's 2022-2023 James R. Killian Jr. Faculty Achievement Award, the highest honor the Institute faculty can bestow upon one of its members each academic year. The award citation credits Peter's "seminal contributions that have forever shaped the foundations of quantum computing. Indeed, quantum computing exists today, in practice, because of Peter Shor." As of 2020, Shor is a Member of the National Academy of Engineering, and in 2022 Fellow of the AMS. Prof. Shor also won the Lise Meitner Distinguished Lecture and Medal in 2022, and the Breakthrough Prize in Fundamental Physics in 2023.
Plenary Talks
Nilanjana Datta (University of Cambridge, England)
Date: Friday (6/27)
Venue: Lydia Mendelssohn Theatre
Title: Quantum Entropies In One-Shot Information Theory And Beyond
Abstract: Entropies play a fundamental role in information theory, with the quantum relative entropy serving as a parent to several key quantities in quantum information theory, namely the von Neumann entropy, quantum conditional entropy, and quantum mutual information. These entropies have important operational interpretations, particularly as optimal rates for various information-theoretic tasks in the “asymptotic memoryless setting”, where resources are assumed to be independent and identically distributed across an arbitrarily large number of uses. In the more realistic “one-shot” setting, these assumptions are lifted. In this talk, we explore two relative entropies, namely, the min- and max-relative entropies, which act as parents to entropies which govern the fundamental limits of information-theoretic tasks in the one-shot setting. These relative entropies are themselves obtainable from two families of relative entropies (the Petz Renyi divergence and the sandwiched Renyi divergence) which arise in the characterization of error exponents in hypothesis testing. The talk will highlight the operational significance of many of the above-mentioned entropies, and culminate in the introduction of a “grandparent entropy” (the α-z relative Renyi entropy) that heads a large family tree of entropies, thus providing a unifying mathematical framework for the analysis of these quantities.
Bio: Nilanjana Datta received the Ph.D. degree in mathematical physics from ETH Zurich, Switzerland, in 1996. From 1997 to 2000, she was a Post-Doctoral Researcher with the Dublin Institute of Advanced Studies, C.N.R.S. Marseille, and EPFL, Lausanne. In 2001, she joined the University of Cambridge, U.K., as a Lecturer in Mathematics with the Pembroke College, and a member of the Statistical Laboratory, Centre for Mathematical Sciences. She is currently a Professor in Quantum Information Theory with the Department of Applied Mathematics and Theoretical Physics, University of Cambridge. Her scientific research interests include quantum information theory and mathematical physics. She is a Fellow of Pembroke College.
Alexander Rakhlin (MIT, USA)
Date: Monday (6/23)
Venue: Lydia Mendelssohn Theatre
Title: On the Foundations of Interactive Decision Making
Abstract: Machine learning methods are increasingly deployed in interactive environments, ranging from dynamic treatment strategies in medicine to fine-tuning of LLMs using reinforcement learning. In these settings, the learning agent interacts with the environment to collect data and necessarily faces an exploration-exploitation dilemma. We present a general framework for interactive decision making that subsumes multi-armed bandits, contextual bandits, structured bandits, and reinforcement learning. We focus on both the statistical aspect of learning—aiming to develop a tight characterization of sample complexity in terms of properties of the class of models—and on the basic algorithmic primitives.
Bio: Alexander Rakhlin is a Professor in the Department of Brain and Cognitive Sciences and the Statistics and Data Science Center at MIT. His research interests lie at the interface of Machine Learning and Statistics, with a focus on online prediction, decision-making, and the theoretical underpinnings of modern learning systems.
Éva Tardos (Cornell University, USA)
Date: Wednesday (6/25)
Venue: Lydia Mendelssohn Theatre
Title: Learning in Strategic Queuing
Abstract: Over the last two decades we have developed good understanding how to quantify the impact of strategic user behavior on outcomes in many games (including traffic routing and online auctions) and showed that the resulting bounds extend to repeated games assuming players use a form of learning (no-regret learning) to adapt to the environment. In this talk, we will focus on repeated interactions that have carry-over effects between rounds: when outcomes in one round affect the game in the future, as in repeated auctions with budgets, as well as queuing systems. In this talk, we study this phenomenon in the context of a game modeling queuing systems: routers compete for servers, where packets that do not get served need to be resent, resulting in a system where the number of packets at each round depends on the success of the routers in the previous rounds. We study the required excess server capacity needed to guarantee that all packets get served in two different queuing systems (with or without buffers) despite the selfish (myopic) behavior of the participants.
Bio: Éva Tardos is a Jacob Gould Schurman Professor of Computer Science, was chair of the Department of Computer Science 2006-2010 and 2020-2023. She was Interim Dean for Computing and Information Sciences 2012-2013 and was Associate Dean for Diversity & Inclusion 2019-2020 at Cornell University. She received her BA and PhD from Eötvös University in Budapest. Tardos’s research interest is algorithms and interface of algorithms and incentives. She is most known for her work on network-flow algorithms and quantifying the efficiency of selfish routing. She has been elected to the National Academy of Engineering, the National Academy of Sciences, the American Philosophical Society, the American Academy of Arts and Sciences, and to the Hungarian and Austrian Academies of Sciences. She is the recipient of a number of fellowships and awards including the Packard Fellowship, the Gödel Prize, Dantzig Prize, Fulkerson Prize, ETACS prize, and the IEEE von Neumann Medal. She co-wrote the widely used textbook Algorithms Design. She has been Editor-in-Chief of the Journal of the ACM and of the SIAM Journal of Computing, and has been editor of several other journals, and was program committee member and chair for several ACM and IEEE conferences in her area.
En-Hui Yang (University of Waterloo, Canada)
Date: Tuesday (6/24)
Venue: Lydia Mendelssohn Theatre
Title: Information Theory Inspired Deep Learning
Abstract: While deep learning (DL) has achieved remarkable empirical success, our understanding of why deep neural networks (DNNs) trained through DL work so well remains incomplete. This talk explores how concepts and tools from information theory (IT) can provide a powerful framework for understanding, analyzing, and improving DL architectures and algorithms. Drawing an analogy between human students and DNNs, we begin by abstracting a classification DNN as a high-dimensional nonlinear function that maps inputs to probability distributions. We then introduce an information geometry perspective to evaluate the structural mapping performance of the DNN—using new metrics that go beyond traditional error probability. Specifically, we use conditional mutual information (CMI) to quantify intra-class concentration in the output distribution space, and introduce a novel quantity, inter-class cross entropy, to measure inter-class separation. Building on these structural metrics, we apply optimization techniques from IT—such as rate-distortion theory and channel capacity—to guide DNN training. This leads to new DL paradigms: CMI-constrained DL (CMIC-DL), knowledge distillation (KD)-resistant DL for protecting model intellectual property, and KD-amplifying DL for enhancing student model performance. Extensive experiments show that these IT-inspired methods outperform state-of-the-art models in accuracy, robustness, and interpretability. The talk concludes by revealing a deep connection between IT and transformer decoder-based large language models, tracing the lineage back to Shannon’s seminal 1948 paper.
Bio: Since 1997, he has been with the Dept. of Electrical and Computer Engineering, University of Waterloo, Ontario, Canada, where he is now a University Professor and former Canada Research Chair in information theory and applications. He is the founding director of the Leitch-University of Waterloo multimedia communications lab, a co-founder of SlipStream Data Inc. (now a subsidiary of BlackBerry), and the founder of BicDroid Inc. Dr. Yang is a Fellow of IEEE, the Canadian Academy of Engineering, and the Royal Society of Canada. He is also a recipient of several awards, including the 2023 Canadian Award for Telecommunications Research; the 2021 IEEE Eric E. Sumner Award; the prestigious Inaugural Ontario Premier’s Catalyst Award in 2007 for the Innovator of the Year；the 2007 Ernest C. Manning Award of Distinction, one of the Canada's most prestigious innovation prizes; the 2013 CPAC Professional Achievement Award; the 2014 IEEE Information Theory Society Padovani Lecture Award; and the 2014 FCCP Education Foundation Award of Merit. Products based on his early inventions and commercialized by his previous company, SlipStream, received the 2006 Ontario Global Traders Provincial Award. His research work has benefited people over 170 countries through commercialized products, video coding open sources, and video coding standards. In 2011, he was selected for inclusion in Canadian Who’s Who.