U.K.-U.S. prize challenges
At Summit for Democracy, the United Kingdom and the United States Announce Winners of Challenge to Drive Innovation in Privacy-Enhancing Technologies That Reinforce Democratic Values.
Transforming financial crime prevention and boosting pandemic response capabilities through privacy-preserving federated learning
Click here to watch the winners being announced at the Summit for Democracy.
The winning solutions combined different PETs to allow the AI models to learn to make better predictions without exposing any sensitive data.
World-leading experts competed for cash prizes from a combined UK-U.S. prize pool of £1.3 m / $1.6m.
The prizes encouraged the development of innovative solutions that address practical data privacy concerns in real world scenarios.
Further information is available in the press release.
Prize challenges have now closed.
Innovating to tackle global challenges
Privacy-enhancing technologies (PETs) have the potential to help us devise data-driven, innovative solutions to tackle the most pressing global societal challenges we're facing, while preserving citizens’ fundamental right to privacy, which constitutes a foundation for democratic societies.
By enabling organisations to share and collaboratively analyse sensitive data in a privacy-preserving manner, PETs open up unprecedented opportunities to harness the power of data through innovative and trustworthy applications.
The United States and United Kingdom will continue to build on their shared interest in advancing responsible innovation in PETs. In May, a joint Demo Day will be held in London to deepen transatlantic communities of practice among UK and U.S. privacy researchers and government representatives. Further collaboration in this space, such as developing tools and guidance to assist practitioners to adopt these technologies effectively and responsibly, is being actively explored.
The goals of the challenges
Drive innovation in the technological development and application of novel PETs
Deliver strong end-to-end privacy guarantees against a set of common threats and privacy attacks, leveraging a combination of input and output privacy techniques
Develop a privacy-preserving solution that is capable of efficiently generating high-utility machine learning models for one of two predefined use-cases in finance and public health, detailed below
The UK and U.S winners of the prize challenges are listed below:
First place (joint): University of Cambridge
First place (joint): STARLIT (Privitar, University College London, Cardiff University)
Third place: Faculty
Fourth place: Featurespace
Special recognition prizes:
University of Liverpool
Red team winner:
White paper prizes:
Faculty, Featurespace, STARLIT (Privitar, University College London, Cardiff University), University of Cambridge, University of Liverpool, DeepMind and OpenMined*, Corvus Research Limited, Diagonal Works, GMV, Privately SA
(*DeepMind and OpenMined chose not to accept any prize funds for this challenge.)
Track A: Financial Crime Prevention
Scarlet Pets (Rutgers University)
PPML Huskies (University of Washington Tacoma, Delft University of Technology, University of Brasilia)
ILLIDAN Lab (Michigan State University, University of Calgary)
Track B: Pandemic Response and Forecasting
puffle (Carnegie Mellon University)
MusCAT (Broad Institute, MIT, Harvard Business School, University of Texas Austin, University of Toronto)
ZS_RDE_AI (ZS Associates)
Red Team Winners:
ETH SRI (ETH Zurich)
Entmoot (Independent researcher)
White Paper Prizes:
MusCAT (Broad Institute, MIT, Harvard Business School, University of Texas Austin, University of Toronto), IBM Research, Secret Computers (Inpher Inc)
Find profiles of the winners from the UK challenge here
Find profiles of the winners from the U.S. challenge here
Structure of the prize challenges
The challenges, which were free to enter, took the form of a multi-stage competition involving a white paper submission, prototype development, and a red-teaming phase.
Participants could select one track or both tracks, or for extra points, develop a solution that works for both.
Transforming financial crime prevention
Innovators were asked to develop solutions that help tackle the challenge of international money laundering, which finances organized crime including human trafficking and terrorist financing, and undermines economic prosperity – costing up to US$2 trillion each year, according to UN estimates.
This illicit activity could be more effectively identified through information sharing and collaborative analytics among financial organisations, but such approaches are made more challenging by legal and technical requirements to ensure customer privacy. Organisations including the Financial Action Task Force have highlighted the potential of PETs to help tackle these barriers by enabling privacy-preserving access to data.
Innovators were asked to develop end-to-end privacy-preserving federated learning solutions to detect potentially anomalous payments, leveraging a combination of input and output privacy techniques. To develop solutions, innovators used synthetic datasets created by Swift, the global provider of secure financial messaging services.
While developing the solutions, innovators in the U.K. were able to engage with the Information Commissioner’s Office (ICO), the U.K. National Economic Crime Centre, and the Financial Conduct Authority (FCA), and innovators in the U.S. were able to engage with the Financial Crimes Enforcement Network (FinCEN).
Forecasting to bolster pandemic response capabilities
Innovators were asked to bolster pandemic response capabilities in both the United States and United Kingdom by developing privacy-preserving federated learning solutions to improve forecasting. The COVID-19 pandemic - which has incurred an immense human cost and socio-economic impact across the globe - demonstrated the importance of preparing for public health emergencies by harnessing the power of data through privacy-preserving data sharing and analytics.
Innovators were asked to develop privacy-preserving federated learning solutions to forecast an individual’s risk of infection, leveraging a combination of input and output privacy techniques. Participants had access to a synthetic dataset created by the University of Virginia’s Biocomplexity Institute, which represented a digital twin of a population with statistical and dynamical properties similar to a real population.
While developing the solutions, innovators in the U.K. were able to engage with the Information Commissioner’s Office (ICO), NHS England, and the UKRI-funded Data and Analytics Research Environments UK (DARE UK), and innovators in the U.S. were able to engage with staff from the Centers for Disease Control and Prevention (CDC).