Apprehending Adversaries Accelerates Agreement: How Accountability Leads to Faster Protocols

Abstract:
Distributed systems have long suffered from malicious behavior and adversarial attacks.  The growth of blockchains and cryptocurrencies has only intensified the problem, especially given the amount of money at stake. A promising response is accountability: the ability of a system to reliably identify participants that misbehave. Accountability can incentivize good behavior, for example, by “slashing” deposits.  Or it can support external enforcement via legal or government bodies.  Overall, accountability can be a powerful tool for deterring disruption and making systems more robust. An interesting question remains: can accountability also make Byzantine Agreement protocols faster and more efficient? In this talk, I will describe two recent results showing that weak forms of accountability can enable faster agreement.  First, I will show how we can iteratively exclude malicious participants, yielding a multi-shot Byzantine Agreement protocol with O(1) average time complexity.  Second, I will discuss how to use unreliable “predictions” (derived from any accountability mechanism) to guide an agreement protocol toward faster outcomes—reducing the time for consensus to O(1) when the predictions are good.  These results suggest that accountability is not merely a tool for robustness, but can also improve performance. Joint work with: Pierre Civit, Naama Ben-David, Muhammad Ayaz Dzulfikar, Faith Ellen, Rachid Guerraoui, Jovan Komatovic, and Manuel Vidigueira.

Short bio:

Seth Gilbert is a Professor of Computer Science at the National University of Singapore, where he works on distributed algorithms, fault tolerance, and the challenges that arise in designing reliable distributed systems. His research spans classical distributed computing, Byzantine agreement, blockchains, and the real-world robustness issues that appear in modern distributed infrastructures. More broadly, Seth’s work focuses on the algorithmic foundations of robustness and scalability, wherever they may arise.

Perspectives on Non-Terrestrial Networks for 6G Communications

Abstract:
This talk presents an overview of the requirements and enabling technologies for 6G, with a focus on non-terrestrial networks (NTN). The talk begins with an introduction to key 6G performance targets, architectural trends, and emerging use cases. It then highlights our long-term research on drone-based networks, which we have been developing since 2018 as a representative NTN use case. I will discuss system design challenges, network control, and communication efficiency in aerial environments. The talk also introduces semantic communication as a promising paradigm for future 6G systems, highlighting how meaning-aware and task-oriented transmission can improve efficiency and robustness in drone and NTN scenarios.

Short bio:
Kaoru Ota received her B.S. and Ph.D. degrees from the University of Aizu, Japan, in 2006 and 2012, respectively, and her M.S. degree from Oklahoma State University, USA, in 2008. She is a Distinguished Professor at the Graduate School of Information Sciences, Tohoku University, Japan, and a Professor at the Center for Computer Science (CCS), Muroran Institute of Technology, Japan, where she served as the founding director. She has been recognized as a Highly Cited Researcher by Clarivate Analytics in 2019, 2021, and 2022, a Fellow of the Engineering Academy of Japan (EAJ) in 2022, and a Fellow of the Asia-Pacific Artificial Intelligence Association (AAIA) in 2025.