Export
Research Overview
Mike Williamson focuses on multi-agent systems, information brokering, and usable security. His early trajectory explored the organizational roles and dynamic behaviors of distributed agents in electronic commerce environments. More recently, his focus has shifted toward empirical security research, specifically analyzing the real-world efficacy and user behavior associated with account recovery mechanisms.
Research Trajectory
Williamson’s earliest observed research interests in the mid-1990s engaged with foundational problems in artificial intelligence and data mining. This period included contributions to the research direction Oversubscription Planning under Resource Constraints, where systems must prioritize objectives under resource constraints, as well as the broader area of Least Commitment Planning. Additional early work intersected with themes such as Hierarchical Structures for Temporal Information Extraction and the area of Pruning Search Trees for High Utility Pattern Mining, reflecting a broad engagement with search efficiency and knowledge representation.
By the late 1990s, Williamson’s focus shifted toward distributed artificial intelligence and the emerging architecture of the internet. Serving as the senior investigator, he explored how autonomous entities could effectively locate and exchange data in open, distributed environments. In Modeling Information Agents: Advertisement, Organizational Roles, and Dynamic Behavior, he and his collaborators formalized how information agents advertise their capabilities and adapt their organizational roles dynamically to meet system demands.
Building on this foundation, Williamson led the development of frameworks for intermediary entities in electronic commerce. The 1997 paper Middle-Agents for the Internet categorized the distinct roles of "middle-agents"—such as matchmakers, blackboard agents, and brokers—that facilitate connections between information providers and requesters. This work analyzed the performance trade-offs of these different architectures, particularly regarding privacy, robustness, and adaptability in large-scale multi-agent systems.
Following a gap in the public publication record, Williamson’s later work transitioned into the domain of usable security and applied cryptography at scale. As a senior author, he contributed to a major empirical analysis of authentication fallback mechanisms in Secrets, Lies, and Account Recovery: Lessons from the Use of Personal Knowledge Questions at Google. This 2015 study evaluated a massive real-world dataset to demonstrate that personal knowledge questions offer significantly lower security than user-chosen passwords. The research highlighted a critical behavioral vulnerability: users frequently provide false answers to security questions in an attempt to make them harder to guess, which paradoxically decreases overall account security and memorability.
Current Directions
The most recent work in the evidence, dating to 2015, investigates the intersection of human behavior and system security at an industrial scale. This direction focuses on empirical evaluations of account recovery systems, challenging long-held assumptions about the efficacy of personal knowledge questions. By analyzing large-scale user data, this research highlights the tension between memorability and security, emphasizing how user attempts to outsmart security mechanisms can inadvertently weaken them. The overarching theme of this recent work is the need for data-driven, usable security protocols that account for actual human behavior rather than theoretical threat models.
Profile updated
Loading new papers…
Loading research neighbors…
See something wrong? Report a mistake