The PETs Prize Challenges: a bilateral collaboration to empower and protect
The PETs Prize Challenges were launched in July 2022 to drive innovation in PETs. The challenges address global challenges such as reducing financial crime and responding to public health emergencies.
PETs are a powerful set of emerging technologies as they can harness the power of data in a secure way enabling cross sector collaboration to solve our world’s greatest challenges. That is why the UK and US governments partnered to deliver this set of prize challenges. The challenges have supported innovators in developing solutions that combine PETs to unlock the economic, scientific and societal benefits of data collaboration.
Beyond this, the challenges have strengthened relations between the two countries to support development of ‘democracy affirming technologies’ that reflect democratic values such as privacy and enable innovation that keeps everyone’s data secure.
An overview of PETs and the prize challenges technology
PETs are technologies which allow data analysis without revealing sensitive information. The PETs Prize Challenges focused on advancing privacy-preserving federated learning, a technique that enables organisations to train machine learning models on sensitive data whilst providing improved privacy protections.
There were three phases to the prize challenge:
- in phase one innovators produced a concept paper, proposing their privacy-preserving solutions. Our phase one winners were announced on 10 November 2022. Read about our phase one innovators: privacy-enhancing technologies prize challenges
- in phase two innovators, termed ‘blue teams’ focused on solution development and submitted code for their solutions on a testing platform hosted by DrivenData which ran federated training and evaluation
- in phase three independent ‘red teams’ were tasked with designing and executing privacy attacks on the solutions developed by blue teams. This was to confirm that sensitive data could not be reconstructed from teams’ solutions. The winning teams were selected based on attacks by red teams and evaluation by an independent panel of PETs experts from government, academia, and industry
Blue teams could focus on one or both of two tracks: financial crime prevention and pandemic forecasting tracks.
Financial crime: in this track teams were tasked with developing solutions to detect potentially anomalous payments, while preserving the privacy of sensitive financial information. Teams were provided with artificially generated or synthetic datasets created by SWIFT in collaboration with BNY Mellon, Deutsche Bank, and MOSTLY AI.
Pandemic forecasting: in this track teams developed solutions to forecast an individual’s risk of infection during a pandemic while preserving the privacy of sensitive health and mobility data. Teams were provided with a synthetic dataset created by the University of Virginia’s Biocomplexity Institute.
The importance of the PETs Challenge: collaboration to navigate global complexity
The use of PETs can advance innovation and collaboration in new fields and help tackle some of our most pressing societal challenges. As data privacy regulations are introduced across the globe, PETs could prove to be vital technologies in navigating an increasingly complex regulatory landscape.
The PETs Prize Challenges have been an excellent demonstration of the power of international collaboration. As a joint initiative between the UK and the US, the PETs Prize Challenges have involved multiple government agencies, regulatory bodies, organisations and assessors on both sides of the Atlantic. This includes:
- Centre for Data Ethics and Innovation
- Innovate UK
- National Institute of Standards and Technology
- National Science Foundation
- White House Office of Science and Technology Policy
- NHS England
- Data and Analytic Research Environments
- UK Information Commissioner’s Office
- UK Financial Conduct Authority
- UK National Economic Crime Centre
- US Financial Crimes Enforcement Network
- University of Virginia Biocomplexity Institute
Introducing our PETs Challenge prize winners
The blue team winners are:
Joint first place
In joint first place: STARLIT, and University of Cambridge
Challenge track: financial crime
Meet the team
STARLIT (Privitar, University College London, and Cardiff University)
Privitar empowers organisations to use their data safely and ethically. Their modern data provisioning solution builds collaborative workflows and policy-based data privacy and access controls into data operations. Privitar combines technology, regulatory expertise, and best practices to support modern data innovation initiatives while navigating regulations and protecting customer trust.
University of Cambridge – the team from the University of Cambridge comes from CaMLSys – the Cambridge Machine Learning Systems lab, based in the Computer Science and Technology department. The team is composed of Professor Nicholas D. Lane, Senior Research Associate, Pedro Porto Buarque de Gusmão, and PhD and Masters students including:
- Xinchi Qiu
- Heng Pan
- Wanru Zhao
- Chenyang Ma
- Filip Svoboda
In third place: Faculty
Challenge track: pandemic forecasting
Meet the team
Faculty is a founder-led company. Since the beginning, their mission has been to bring the benefits of artificial intelligence to everyone. They have worked on some of the biggest and most difficult challenges faced by major organisations, using ‘decision intelligence’ to help organisations make better decisions on the things that matter.
In fourth place: Featurespace
Challenge track: financial crime
Meet the team
For nearly a decade, Featurespace has been at the forefront of modernising financial crime prevention in financial services institutions around the world. Their next generation machine learning models provide industry leading predictive performance.
The red team winner is: Trūata
Meet the team
Trūata is a PET company that specialises in quantifying the privacy risk in datasets and ensuring analytical outputs meet required privacy thresholds. Trūata’s red team for this challenge contains six data scientists with specialties in data privacy, privacy engineering, machine learning and statistical analysis. Along with a vast amount of experience in executing adversarial attacks on outputs created using privacy-enhancing technologies.
Our panel of assessors also awarded several special recognition prizes for solutions that demonstrated excellence in a particular theme:
Usability prize (for a solution that is deemed particularly user-friendly):
- Featurespace – awarded for in-depth exploration of the impact of privacy-parameter tuning on the system
- Faculty – awarded for in-depth exploration of the impact of privacy-parameter tuning on the system
Novel modelling and design prize (for solutions with particularly innovative approaches to the modelling task and system architecture):
- Privitar – awarded for novel system design using private set intersection and a flag aggregator
- University of Liverpool – awarded for novel privacy architecture using hierarchical privacy domains
Regulatory distinction prize for a team that can most effectively demonstrate compliance with relevant regulatory frameworks and principles, as decided by our independent panel of assessors (this includes data protection, and finance and healthcare regulation):
- Diagonal – awarded for effectively demonstrating meaningful consideration of the regulatory implications of their solution and its consequences for information governance
The second Summit for Democracy
The PETs Prize Challenges UK and US winners were profiled on 30 March 2023 at the second Summit for Democracy. The summit is convened by the US and co-hosted with the governments of Costa Rica, the Netherlands, Republic of Korea, and Republic of Zambia.
The aim of the summit is to:
- champion a vision of our world that is grounded in democratic values and supports transparent, responsive, accountable governance, rule of law, respect for human rights, and a prosperous future for all
Congratulations to the winners and participants for all their hard work in what has been an exciting competition!
A special thank you to our assessors for their time and input in a competitive judging process.
You can connect with Sarah on LinkedIn
Top image: Credit: MicroStockHub, iStock, Getty Images Plus via Getty Images