DRAG*N
Research Projects
This page lists current funded research projects being led by members of the Data Research, Access, and Governance Network (DRAGoN)
GRAIMatter: Guidelines and Resources for Artificial Intelligence Model Access from Trusted Research Environments
Funder: UK Research And Innovation - DARE UK
Duration: January 2022 – August 2022
Duration: January 2022 – August 2022
Trusted research environments (TREs) provide a secure location for researchers to analyse data for projects in the public interest – for example, to provide information to the Scientific Advisory Group for Emergencies (SAGE) to fight the COVID-19 pandemic. TRE staff check outputs to prevent disclosure of individuals’ confidential data.TREs have historically supported only traditional statistical data analysis, and there is an increasing need to also facilitate the training of artificial intelligence (AI) models. AI models have many valuable applications, such as spotting human errors, streamlining processes, helping with repetitive tasks and supporting clinical decision making. The trained models then need to be exported from TREs for use.
The size and complexity of AI models presents significant challenges for the output checking process. Models may be susceptible to external hacking: complicated methods to reverse engineer the learning process to find out about the data used for training, with more potential to lead to re-identification than conventional statistical methods.
With input from public representatives, GRAIMatter will assess a range of tools and methods to support TREs to assess output from AI methods for potentially identifiable information, investigate the legal and ethical implications and controls, and produce a set of guidelines and recommendations to support all TREs with export controls of AI algorithms.
The size and complexity of AI models presents significant challenges for the output checking process. Models may be susceptible to external hacking: complicated methods to reverse engineer the learning process to find out about the data used for training, with more potential to lead to re-identification than conventional statistical methods.
With input from public representatives, GRAIMatter will assess a range of tools and methods to support TREs to assess output from AI methods for potentially identifiable information, investigate the legal and ethical implications and controls, and produce a set of guidelines and recommendations to support all TREs with export controls of AI algorithms.
PI: Professor Emily Jefferson, University of Dundee
Contact: [email protected], [email protected], & [email protected] (Co-Is)
Contact: [email protected], [email protected], & [email protected] (Co-Is)
The Value of Data Governance
Funder: Centre for Agriculture and Bioscience International (CABI)/Open Data Institute (ODI)
Duration: July 2019 – December 2020
Duration: July 2019 – December 2020
In July 2019, DRAG*N economists were commissioned by CABI/ODI to measure the value of improving data governance and access in Gates Foundation Programmes. The initial aim of this study was to develop, and then assess, the applicability of a formal framework to evaluate the value of data governance, with the aim of providing CABI and its partners with a standardised approach for developing and evaluating any future investment. The second aim was to apply the framework to evaluate the value of data governance in a specific CABI funded project. The expectation was that the project evaluated would not only benefit through gaining a greater understanding of the value of data governance within their project, but also gain an understanding of how this was created. The case study selected was the Supporting Soil Health interventions in Ethiopia.
Following an initial round of desk research, two rounds of primary data collection have ensued. Initially a number of semi structured interviews with members of the project team where undertaken in order to identify the key themes and mechanisms. Subsequently the findings from the qualitative research have been quantitatively analysed through a broader survey of expert stakeholders. The final report is expected to be published by the end of 2020.
Following an initial round of desk research, two rounds of primary data collection have ensued. Initially a number of semi structured interviews with members of the project team where undertaken in order to identify the key themes and mechanisms. Subsequently the findings from the qualitative research have been quantitatively analysed through a broader survey of expert stakeholders. The final report is expected to be published by the end of 2020.
Contact: [email protected]
Wage and Employment Dynamics
Funder: ADR UK
Duration: October 2019 – March 2022
Duration: October 2019 – March 2022
This project links data from various official surveys and administrative datasets, with the objective of providing new insights into the dynamics of earnings and employment in the UK. The project aims to create the basis for a sustainable, documented ‘wage and employment spine’ with the potential to fundamentally transform UK research and policy analysis across a vast range of topics. Alongside the creation of data infrastructure, the project will also generate research findings of direct interest to policy makers. Public benefit will be maximised through the provision of high-quality metadata and training for users.
The project is led by Felix Ritchie and Damian Whittard (UWE) and with co-investigators from UCL (Prof Alex Bryson), CASS Business School (John Forth) and NIESR (Lucy Stokes). The project is being supported by Arusha McKenzie and Van Phan (UWE).
The project is led by Felix Ritchie and Damian Whittard (UWE) and with co-investigators from UCL (Prof Alex Bryson), CASS Business School (John Forth) and NIESR (Lucy Stokes). The project is being supported by Arusha McKenzie and Van Phan (UWE).
Automated disclosure control for research outputs
Funder: Eurostat
Duration: January 2020 – December 2020
Duration: January 2020 – December 2020
Confidential microdata is increasingly made available for research purposes to accredited researchers in secure environments. In all these environments, there is a residual risk that statistical findings could inadvertently breach confidentiality, and so outputs are always reviewed by trained staff before publication. This is time consuming, and many outputs can be cleared (or not) with simple rules that can be applied automatically, but to date there has been no general solution.
Felix Ritchie and Elizabeth Green (Economics) and Jim Smith (Computing) have been commissioned to develop proof-of-concept of a general solution. Although developed in Stata, the proof-of-concept will show a path to a non-proprietary solution, as well as addressing the key requirement of user and manager acceptability
Felix Ritchie and Elizabeth Green (Economics) and Jim Smith (Computing) have been commissioned to develop proof-of-concept of a general solution. Although developed in Stata, the proof-of-concept will show a path to a non-proprietary solution, as well as addressing the key requirement of user and manager acceptability
Contact: [email protected]
Advanced Ethical Models of Data Governance
Funder: UWE Bristol (HAS/FBL Faculties)
Duration: September 2020 – August 2021
Duration: September 2020 – August 2021
Recent scientific and technological advancements (AI, big data) are substantially modifying contemporary ethics. Essential tenets of human morality such as individual responsibility, free will, and personal power, are losing their centrality due to the growing impact of artificial agents on today’s reality. This resulted in new models of distributed morality and big data-induced hyper-networked ethics that are currently being designed and tested. This paradigm shift has consequences in data governance insofar as it alters the private/public divide and the role of collective and societal components for calculating risks and costs of data access.
We propose a cross-faculty project that aims at gathering experts from various disciplines (ethics, economics, business and law) to discuss and produce key deliverables in this research area. Our fundamental task is to identify and test new ethical approaches and models that can be successfully applied to data governance. To do that we will first (1) conduct doctrinal research based on recent developments in information and data ethics and will then (2) analyse the wider contextual constraints that the implementation of such models involves. Francesco Tava (HAS) will be responsible for 1, while Elizabeth Green (FBL) will be responsible for 2. Felix Ritchie (FBL) will provide specialist advice throughout the whole research regarding the contextual analysis of the difficulties and constrains that implementing ethical models in concrete scenarios involves, particularly institutional factors, and will contribute to the elaboration of future research bids. Both co-applicants will commit to apply for an extensive external grant (Wellcome Trust Seed Award) in order to further develop the project.
We propose a cross-faculty project that aims at gathering experts from various disciplines (ethics, economics, business and law) to discuss and produce key deliverables in this research area. Our fundamental task is to identify and test new ethical approaches and models that can be successfully applied to data governance. To do that we will first (1) conduct doctrinal research based on recent developments in information and data ethics and will then (2) analyse the wider contextual constraints that the implementation of such models involves. Francesco Tava (HAS) will be responsible for 1, while Elizabeth Green (FBL) will be responsible for 2. Felix Ritchie (FBL) will provide specialist advice throughout the whole research regarding the contextual analysis of the difficulties and constrains that implementing ethical models in concrete scenarios involves, particularly institutional factors, and will contribute to the elaboration of future research bids. Both co-applicants will commit to apply for an extensive external grant (Wellcome Trust Seed Award) in order to further develop the project.
Contacts: [email protected]; [email protected]
Autumn school in data governance for low- and middle-income countries (LMICs)
Funder: National Institute for Health Research
Duration: October 2020 – December 2020
Duration: October 2020 – December 2020
The DRAG*N team have developed training in data governance for a wide range of organisations across the world, including the accredited training required under the UK Digital Economy Act 2017, and the self-study material for users of Eurostat data. Unlike more traditional training courses, these are based on EDRU (evidence-based, default-open, risk-managed, user-centred) principles and psychological models of human interaction.
NIHR commissioned Julie Mytton (project lead, Faculty of Health and Applied Sciences), Felix Ritchie and Elizabeth Green (Faculty of Business and Law) to develop a residential ‘summer school’ in data governance targeted and NIHR research units/groups in LMICS/. With the intervention of Covid-19 this has now become an ‘autumn school’ running virtually throughout Autumn 2020.
NIHR commissioned Julie Mytton (project lead, Faculty of Health and Applied Sciences), Felix Ritchie and Elizabeth Green (Faculty of Business and Law) to develop a residential ‘summer school’ in data governance targeted and NIHR research units/groups in LMICS/. With the intervention of Covid-19 this has now become an ‘autumn school’ running virtually throughout Autumn 2020.
Contact: [email protected]
Process evaluation of R&D and innovation in data access and governance
Funder: Open Data Institute (theodi.org)
Duration: October 2020 – March 2021
Duration: October 2020 – March 2021
An interdisciplinary team from DRAG*N are undertaking a process evaluation of the Open Data Institute’s Research and Development programme. The ODI four-year £8million R&D project was funded by Innovate UK to support and build upon the UK’s strength in data access. As part of the project the UWE team will review the methodology of placing an economic value on data access and good governance, building on the findings of the CABI project (above).
The UWE interdisciplinary team consists of specialist in data governance, operations management, economic evaluation, psychology and ethics. Their study will help shape future ODI strategic and operational development, enabling them to further their mission to support an open culture: a data infrastructure that is as open as possible; data literacy and capability for all; and open innovation. The project is led by Damian Whittard and Kyle Alves.
The UWE interdisciplinary team consists of specialist in data governance, operations management, economic evaluation, psychology and ethics. Their study will help shape future ODI strategic and operational development, enabling them to further their mission to support an open culture: a data infrastructure that is as open as possible; data literacy and capability for all; and open innovation. The project is led by Damian Whittard and Kyle Alves.
Contact: [email protected]
