This special issue of the IEEE Journal on Selected Areas in Information Theory is dedicated to the memory of Toby Berger, one of the most important information theorists of our time, who passed away in 2022 at the age of 81. He made foundational contributions to a wide range of areas in information theory, including rate-distortion theory, network information theory, quantum information theory, and bio-information theory. He also left a deep imprint on diverse fields in applied mathematics and theoretical engineering, such as Markov random fields, group testing, multiple access theory, and detection and estimation. Well known for his technical brilliance, he tackled many challenging problems, but above all, it is his pursuit of elegance in research and writing that shines throughout his work. The goal of this special issue is to celebrate Toby Berger’s lasting legacy and his impact on information theory and beyond. Original research papers on topics within the realm of his scientific investigations and their “offspring”, as well as expository articles that survey his pioneering contributions and their modern developments, are invited.

Over the past decade, machine learning (ML), that is the process of enabling computing systems to take data and churn out decisions, has been enabling tremendously exciting technologies. Such technologies can assist humans in making a variety of decisions by processing complex data to identify patterns, detect anomalies, and make inferences. At the same time, these automated decision-making systems raise questions about security and privacy of user data that drive ML, fairness of the decisions, and reliability of automated systems to make complex decisions that can affect humans in significant ways. In short, how can ML models be deployed in a responsible and trustworthy manner that ensures fair and reliable decision-making? This requires ensuring that the entire ML pipeline assures security, reliability, robustness, fairness, and privacy. Information theory can shed light on each of these challenges by providing a rigorous framework to not only quantify these desirata but also rigorously evaluate and provide assurances. From its beginnings, information theory has been devoted to a theoretical understanding of the limits of engineered systems. As such, it is a vital tool in guiding machine learning advances.

Modern networks rely on a variety of technologies to sense the environment for static or locomotive objects, in particular their shapes, distances, directions, or velocities. Sensing is a key feature in these networks and enables for example autonomous driving, motion sensing in health applications, target detection in smart cities, or optimal beam selections in millimeter wave communication. Besides these exciting new applications, sensing remains an important feature also for traditional applications such as temperature monitoring, or earthquake or fire detection, where new technologies are exploited including continuous feature monitoring over the entire range of an optical fiber network. The purpose of this special issue is to report on new exciting applications of sensing in modern networks, novel sensing architectures, innovative signal processing mechanisms related to sensing, as well as new results on the fundamental performance limits (resolution, sample complexity, robustness) of sensing systems. Particular focus will be on joint systems that integrate sensing with other tasks, for example communication, information retrieval (estimation, feature extraction, localization), super-resolution.

Causal determinism, is deeply ingrained with our ability to understand the physical sciences and their explanatory ambitions. Besides understanding phenomena, identifying causal networks is important for effective policy design in nearly any avenue of interest, be it epidemiology, financial regulation, management of climate, etc. In recent years, many approaches to causal discovery have been proposed predominantly for two settings: a) for independent and identically distributed data and b) time series data. Furthermore, causality- inspired machine learning which harnesses ideas from causality to improve areas such as transfer learning, reinforcement learning, imitation learning, etc is attracting more and more interest in the research community. Yet fundamental problems in causal discovery such as how to deal with latent confounders, improve sample and computational complexity, and robustness remain open for the most part. This special issue aims at reporting progress in fundamental theoretical and algorithmic limits of causal discovery, impact of causal discovery on other machine learning tasks, and its applications in sciences and engineering.

This special issue of the IEEE Journal on Selected Areas in Information Theory is dedicated to the memory of Alexander Vardy, a pioneer in the theory and practice of channel coding. His ground-breaking contributions ranged from unexpected solutions of longstanding theoretical conjectures to ingenious decoding algorithms that broke seemingly insurmountable barriers to code performance. Inspired not just by the mathematical beauty of coding theory but also by its engineering utility, Alexander Vardy developed novel coding techniques that have had a profound impact on modern information technology, including computer memories, data storage systems, satellite communications, and wireless cellular networks. At the same time, his innovations left their imprint on other scientific disciplines, such as information theory, computer science, and discrete mathematics.