Establish a joint Centre for Modelling Futures

A Centre for Modelling Futures would combine expertise across disciplines and methodologies to advance computational modelling for complex decision making in policy and business. It responds to the recommendation in the recent ‘Blackett Review’ on Computational Modelling (GO-Science, 2018), chaired by Sir Mark Walport, which highlighted the need to step up the UK’s modelling effort, integrate it better into the policy process and expand existing expertise. The report proposes the establishment of a centre of expertise for the public and private sectors to promote exchange of expertise and independent critique of models.

Through a comprehensive programme of scholarship, applied research and training, the Centre would:

1. Develop modelling's conceptual foundations and establish modelling as a recognised method for policy development in the private and public spheres.

2. Facilitate a significant uptake in the use of computational models through the development and integration of modelling methods and methodologies drawing from data science and artificial intelligence research.

3. Advance the usability and applicability of computational models by incorporating existing and new data sources in novel ways and developing more effective visualisations.

4. Entrench modelling within the policy process by close collaboration with stakeholders and users and through the ongoing provision of services and expertise.

5. Show how modelling can involve citizens and enhance democratic accountability.

6. Build capacity in modelling through the development of a coherent curriculum across academic levels and the provision of specialist short courses and summer schools.

The Centre's work would show how modelling by drawing on AI techniques can respond to a set of challenges:

* Complexity: can models be used to understand better the complexity of society?

* Dynamics: can the inherently dynamic nature of society be represented though models?

* Micro-macro: can the two-way interactions between structure and agency be modelled using ideas of emergence?

* Uncertainty: can models help deal with situations of radical uncertainty?

* Time and space: can models represent the consequences of time and space on places, individuals and institutions?

* Forecasting: can modelling generate scenarios that are useful for exploring possible futures?

The Centre would be set up as a consortium of business, government and academic partners with an initial lifespan of 5 years, funded through UKRI and business.

Why the contribution is important

Designing the present to achieve the intended future is the main goal of planning and policy making, of good governance. In many advanced economies tackling the future seems to revolve around managing change, often conceptualised as risks, to retain stability. These risks run through all areas of society: radicalised opinions and political stability, natural resources and consumption, national security, migration etc. Many of these risks are complex, involving many individuals, interdependent systems, and emergent consequences that are often impossible to predict. The Centre for Modelling Futures will promote computational modelling as a method for good governance of complex systems, allowing for the exploration of scenarios, creating futures to shape the present.

While models have been an important part of the toolkit of policy analysts in government and business for many decades, the advent of powerful computer technologies and the development of artificial intelligence techniques such as machine learning and multi-agent systems has transformed the potential of models and the practice of modelling — first in engineering and now increasingly in social science and policy. Computational modelling allows us to build theories and trial ideas that take account of the dynamic, complex nature of human society in a way that has previously been impossible. Human societies are dynamic systems, made up of many agents, that make decisions and influence each other, co-creating social and societal phenomena such as markets, fashions, norms and institutions, evolving over time. Computational models come in a variety of flavours, but all focus on modelling the dynamic, evolving aspects of systems. ​System models​ provide dynamic representations of systems, by integrating different modelling domains into a coherent single model. They are particularly good at modelling ​flows ​and ​feedback​. ​Agent-based models simulate individual social actors and their interactions with each other and the environment. They are particularly good at modelling ​heterogeneity​ and​ emergence​, essential features of social systems. ​Dynamic social network analysis​ provides models of the evolution of social relationships​.

The Centre would bring together world-class academic leaders in social science and policy modelling, partners from large and small businesses who develop or use computational models, and a range of government departments and agencies who develop and use models to inform public policy. A comprehensive programme of career development would be provided for the researchers and research students who will be recruited into the Centre, and one outcome of the Centre will be a cadre of people trained and experienced in advanced computational modelling. Supplementing the core Centre team, a network of academic and non-academic Fellows would join the Centre for weeks or months. They will offer expert advice on areas that need specialist input. The Centre should become the UK hub around which expertise in applying computational modelling will cluster and to which people come for guidance.

Given the relative advancement of the field in the UK and the high awareness of the potential policy significance of computational modelling, the Centre should expect to become a world-leader, generating high international impact for UK researchers and policy makers.

 

 

 

by Nigel on July 01, 2018 at 02:23PM

Current Rating

0.0
Average score : 0.0
Based on : 0 votes

Comments

  • Posted by OAIteam July 11, 2018 at 13:26

    Thank you for your comment. You are right that future outcomes depend on contemporary considerations – in the context of policy and more widely. Delivering models that pave the way with certainty is a constant challenge, so much so that we have entire industries that address uncertainty – from risk managers to finance and insurance firms. These industries are also developing AI tools to facilitate the provision of their services.

    It sounds like much of what you set out depends on the availability of data. This is something the Office for Artificial Intelligence is working towards in our commitment to deliver data trusts, or agreements whereby data is shared across organisations participating in a given data trust. We intend to pilot the first data trust later this year. In the meantime, we’re working with partners – both within and outside government – on the necessity of data to develop AI tools.

    In addition to modelling complex social systems, AI can be integrated into the delivery of services and integrated into existing processes with a view to gaining efficiencies. This is something we are currently exploring.

    Government is also exploring how AI can be used to generate insights in specific contexts with a view to promoting positive outcomes for the public. The first demonstration of this is the recently launched AI and Data Mission on Early Diagnosis (for more information, follow this link: https://www.gov.uk/[…]/missions). While not a society-wide application, we consider it an opportunity to deploy AI solutions to health so that illnesses can be identified and treatments started earlier with a view to saving the lives of those affected.
Log in or register to add comments and rate ideas
UA-119667974-1