Skip to main content

News

Girton PhD student's article on using machine learning to inform COVID-19 measures in London wins award

Co-Authors of the paper pictured together, L-R: James Walsh and Andrew Wang

Girton College PhD student in Engineering (2021) and MCR President, James Walsh, has won the 2024 Wilkes Award for the paper “Near Real-Time Social Distance Estimation in London”.

The Wilkes Award is presented once a year to the authors of the best paper published in the volume of The Computer Journal from the previous year, based on originality and quality of theme and treatment. The Wilkes Award is named after Sir Maurice Wilkes, who was Director of the Cambridge Computer Laboratory throughout the whole development of stored program computers starting with EDSAC; inventor of labels, macros and microprogramming; with David Wheeler and Stanley Gill, the inventor of a programming system based on subroutines.

The winning paper explores the work of Project Odysseus, a collaboration between the University of Cambridge, the University of Warwick, and The Alan Turing Institute, which provided data about the busyness of London streets to the Greater London Authority (GLA) and Transport for London (TfL).

To mitigate the COVID-19 pandemic, policymakers at the Greater London Authority relied upon prompt, accurate and actionable estimations of lockdown and social distancing policy adherence. Transport for London reports they implemented over 700 interventions such as greater signage and expansion of pedestrian zoning at the height of the pandemic’s first wave with our platform providing key data for those decisions. Large well-defined heterogeneous compositions of pedestrian footfall and physical proximity are difficult to acquire, yet necessary to monitor city-wide busyness and consequently discern actionable policy decisions. 

The project therefore leveraged existing urban air quality machine learning infrastructure to process over 900 camera feeds in near real-time to generate new estimates of social distancing adherence, group detection, and camera stability. In the paper, they introduce a platform for inspecting, calibrating, and improving upon existing methods, describing the active deployment on real-time feeds and providing analysis over 18 months.

According to the paper Transport for London (TfL) made more than 700 interventions in the first wave of the pandemic, such as increased signage and pedestrian zoning to create more space, based on data produced by the project’s machine learning algorithm.

James shared:

“We are thrilled to be honoured with this award. Due to the airport closures at this time, I wound up trapped for months over 5,000 miles away from London. Labelling street furniture via satellite imagery to calibrate the cameras was a large component to how I processed the pandemic so far from home. The later success and positive impact of this project is an even greater reward and is owed to the resourcefulness of the team, the leadership of our research group, and partnership with the Greater London Authority.”

Photograph: James Walsh (Engineering PhD, 2021) with co-author Andrew Wang.