Keynotes

Keynotes | September 22

Title: Cybersecurity in the Age of Artificial Intelligence 

Abstract: The security of the digital ecosystem is a top priority worldwide, as the daily lives of people, public services, and the operations of companies and organizations increasingly rely on digital infrastructures. Discovering weaknesses and vulnerabilities in complex distributed systems made up of diverse software and hardware components is a challenging task. Artificial intelligence (AI) provides a range of approaches for early vulnerability discovery, cyber-threat detection, and incident response. These techniques have been investigated over the past decades and are widely deployed in commercial cybersecurity products. From an attacker’s perspective, AI enables the crafting of more targeted and adaptive attacks that are harder to detect. The effectiveness of AI techniques depends on properly addressing the specific weaknesses and vulnerabilities related to data-driven learning approaches, including poisoning attacks and adversarial manipulation. This talk provides an overview of the role of artificial intelligence in cybersecurity and the pitfalls that must be addressed from an attack–defense perspective.

Title and abstract: to be defined

Keynotes | September 24

Title: Byzantine Eventual Consistency for Reliable Local-first Software

Abstract: Local-first software, a growing industry movement, aims to make collaborative applications (such as text and graphics editors, spreadsheets, etc.) resilient against server failure by making the client software on the end user’s device a full replica of the application state, and allowing it to sync with other replicas in a decentralised fashion. This has many advantages, but also introduces new challenges, since the system can no longer rely on a central server to maintain the consistency and security of the collaborative document, and practical systems have to tolerate potentially malicious clients. This talk gives an overview of our work on Byzantine Eventual Consistency, and how the open source Automerge library makes this approach efficient.