Weak signal detection - a unified conceptual framework

Researchers from the Future Resilient Systems (FRS) programme conducted a comprehensive interdisciplinary review of weak signal detection and propose a unified conceptual framework to improve weak signal exploitation in social systems.

Weak signal detection - a unified conceptual framework

Dr Dionysios Georgiadis and Prof. Martin Raubal from the Future Resilient Systems (FRS) programme propose a unified conceptual framework for weak signal detection in their study "An interdisciplinary review on weak signal detection".

When managing systems that are embedded in complex environments, being able to anticipate and navigate discontinuities can determine the difference between perishing and thriving. To this end, one may actively scan the environment for weak signals: seemingly minor irregularities that may foretell impending discontinuities. This approach has been studied among different areas, from the detection of ball bearing faults in rotating machinery, to the rupture of fuel tanks, the prediction of riots and political uprisings, and financial crises.

Despite the universality of this challenge, each piece of literature has focused on idiosyncratic definitions, tools, and solutions. As a result, there is no unifying framework for exploiting weak signals, and interdisciplinary convergent research is impeded.

Seeking to remedy this issue, Dr Georgiadis and Prof. Raubal take a holistic view on how weak signals are exploited in different fields and propose a unified conceptual framework. Based on this framework, they identified research directions that may enable social sciences to draw methodology from mathematically rigorous fields. This approach stands to drastically improve weak signal exploitation in social systems by reducing human involvement and lowering the detecting threshold for weak signals.

For full details, please download the working paper Download “An interdisciplinary review on weak signal detection” (PDF, 451 KB).

JavaScript has been disabled in your browser