Professor Simon Maskell, University of Liverpool

Research Chairs and Senior Research Fellowships 2024-25

Intelligence, Surveillance, and Reconnaissance (ISR) algorithms use a model for the real world to extract information pertinent to decision-making, which typically relates to many objects. The University of Liverpool’s work for the Defence Science and Technology Laboratory (Dstl) over recent years, spanning land, sea, air and space, has made clear that the existing state-of-the-art struggles to respond adequately to an adversary that is actively attempting to disrupt our decision-making. This is because: models for real-world objects rarely consider deception; solutions are centralised (and so vulnerable to attack), bespoke, and/or suboptimal; automated decision-making fails to account for the long-term impact of short-term decisions, leaving the system unable to capitalise fully on its available resources or heavily reliant on human intervention.

Dstl’s lead in developing the open-source software framework, Stone Soup, provides a conveyor belt between research and application. Professor Simon Maskell and his team’s current research is populating Stone Soup with both next-generation and generation-after-next solutions. However, their projects are relatively short-term and tactical endeavours.

This Research Chair fulfils the need for a longer-term strategic perspective that augments the continuing stream of more tactical projects they assume will persist. The Chair will ensure that the portfolio totals more than the sum of its parts. More specifically, it will ensure that Dstl has algorithms, in Stone Soup, that can: cater for deception; capitalise on machine learning as a generic strategy for developing distributed algorithms that are inherently robust to attack and which explicitly approximate the optimal centralised solution; average over hypothetical ‘roll-outs’ of the future to deduce the optimal short-term decisions with respect to meeting long-term information goals.

---

Personal website

Academic profile

LinkedIn