Probabilistic Programming open call

Category

Research

Closing date

Key dates

10 October 2023: The deadline for submissions is 10 October 2023, 13:00 BST
06 November 2023: The review panel decisions will be shared with applicants for any invitations to "revise and resubmit"
By 30 November 2023: Final decisions will be communicated to all applicants and offer letters/contracts will be issued for signature
01 January 2024 – 01 March 2024: Projects must start between 01 January 2024 and 01 March 2024, as soon as the contracts have been signed

Background

The Alan Turing Institute is the UK’s national centre for AI and data science and is currently developing a portfolio of research supporting emerging Fundamental AI research areas.

Our current Fundamental AI strategy aims to:

  • pump prime promising and emerging areas of research with the aim of growing the UK’s capacity in these areas 
  • support activities that require a national perspective, leveraging the Turing’s role as the national centre

To this end, the Turing is pleased to announce funding for a portfolio of projects around the theme of probabilistic programming.

Probabilistic programming is an emerging research area that has attracted interest from the Artificial Intelligence (AI) and Machine Learning (ML) communities, as well as statistics and theoretical computer science/semantics. The basic concept of probabilistic programming is to use ideas from programming languages to structure complex statistical models. Probabilistic programming has found applications in a range of settings, perhaps most notably in pandemic modelling.

What we are looking for

In April 2023, Turing hosted a workshop on ‘Probabilistic programming: its challenges and opportunities’ where we consulted with the research community to identify research priorities that could move theory into practice and make Probabilistic programming more widely used and understood.

Stemming from the workshop we are now inviting applications for project proposals that   promote the development, adoption, and awareness of probabilistic programming within the UK. We will fund a portfolio of projects with a particular emphasis on the following themes:  

  • Compositionality of probabilistic programs
  • Parallelization of probabilistic programs
  • Programmable inference
  • Semantics of probabilistic programs
  • Verification of probabilistic programs
  • Probabilistic programs and neural nets
  • Flagship applications of probabilistic programming
  • Community projects to support dissemination/adoption of probabilistic programming technologies.

This list is non-exclusive, and we encourage applicants to think creatively about how they might address the requirements of this call.

Please note that you can apply even if you did not attend the workshop. 

Please be aware that collaboration with the Turing community is key throughout the project lifecycle. All project teams selected from this call will be expected to engage with one another and with Turing researchers and staff to actively collaborate and help develop the “probabilistic programming” research theme within the Turing and grow a UK community.  As part of your application, you will need to outline how you will contribute to the strategic development of this theme at the Turing.

This call is aimed at developing areas of work that are formative (i.e., pre-large scale funding stages), which do not fit standard funding calls, or which would benefit from the Turing being a national institute.    

We are expecting to deliver these projects within the Turing, and therefore successful applicants and their teams (as applicable) will be onboarded into the Turing. Researchers will be engaged via a secondment agreement to Turing - see bottom of page to download a copy, which must be signed prior to the project starting, subject to the agreement of your employer. The project teams will have access to Turing resources and platforms and supported by the Turing operational and business teams. If new full-time employees are required for a project (e.g., postdoctoral research associates) they will be recruited and employed by the Turing.

Turing can offer support from our professional teams for your project, such as research engineers and data wranglers, if needed and fully justified by the project’s objectives and methodology. If your project will require engagement of these teams, please contact us in a timely manner and we can advise on costs, availability and scope for engaging members of these teams. 

We anticipate that all project IP will be owned by Turing with the intention that outcomes will be published on an open-source basis under a creative commons license. Turing is committed to sharing algorithms and methods publicly.

Who can apply

We invite researchers from any UK-based universities and research institutes to submit an application. 

The lead applicant or Principal Investigator must be based in a UK university or research institute. The co-investigators do not have to be UK-based or based in a university.

All people named on the proposal must have permission from their organisation to apply to ensure the organisations agree to the Terms and Conditions, given via a Submission Approval Letter - see bottom of page to download a copy. 

Costs we will fund

A total of approximately £2.5M is available for this call. We expect to fund 3 to 5 projects for 24 to 36 months, but applications for smaller projects are welcome. Each proposal will be assessed on its own merits using the assessment criteria (given below) and in relation to the whole suite of submitted applications.  If we receive multiple similar proposals, there may be a suggestion from the review panel to collaborate and develop a joint proposal.

All projects must start between 01 January 2024 – 01 March 2024.

Eligible costs include: 

  • Salary of personnel working directly on the project – this could include, for example, PIs, postdoctoral research associates, research assistants. Full salary and on-costs of seconded researchers will be paid to the university, pro-rated to %FTE and duration of secondment. VAT at the standard 20% rate is applicable to secondments.
  • Salary or costs of technical or professional personnel allocated to the project from the Turing – this can include for example, data managers/wranglers, data scientists or software engineers.  
  • Travel and subsistence for project researchers (e.g., attending conferences, travelling to/from the Turing/other collaborators)  
  • Conference or event attendance fees (where conference/event is directly applicable to the research project)  
  • As a Turing member of staff, you can purchase cloud computing resources which provide services such as virtual machines, databases storage as well as targeted services in fields including analytics, machine learning, and internet of things.
  • Other costs which are specifically justified for the project e.g., books, events or catering costs, specific laptops (laptops will be provided by Turing for researchers 100% FTE seconded to Turing and do not need to be costed in) 
  • Open access publication costs. 

Note that overheads will not be paid as part of the Secondment Agreements.

Requested costs must be pre-approved by the university, so please ensure to include this statement of approval as part of the Submission Approval Letter - see bottom of page to download a copy.

Any Turing-related costs (e.g., computing resources, research engineers, etc) must be discussed and pre-approved by the Turing team. Please contact the programme manager (Ruth Drysdale: [email protected]) in a timely manner to arrange this.

Please note, the budget requested may not be funded to the maximum requested amount. The Turing leads may work with applicants to build collaborative projects which may involve updating costs for projects. 

As secondees, researchers can request to receive access passes to use the Turing office space in the British Library, London and its meeting rooms. 

How to apply

Applications must be submitted via the online portal at (call ID: 12056 Probabilistic Programming) https://ati.flexigrant.com/startapplication.aspx?id=12056  

If you have not already done so you must first register on the Flexigrant system.

If you have any questions regarding the application form or using the online system, please contact the Research Project Manager: Chiara Sbruazzo [email protected]  

Applications must be submitted by 13:00 on 10th October 2023.

In addition to the FlexiGrant form (Appendix 1), the application requires the submission of four documents in PDF format:

1. CVs (up to 2 pages each) of the lead applicant/PI and each of the named researchers, which should clearly demonstrate their track record relevant to the call.

2. 4-page proposal including these sections:

  • Summary of the proposed project (100 words)
  • Project objectives
  • Description of the work you propose to undertake, clearly outlining how it fits to the call, and addressing potential risks and mitigation
  • Expected outputs and impact (e.g.: research results; software; website; datasets; etc)
  • You may include references, if needed, as an additional half page maximum (in addition to the 4-page limit of the proposal)

3. Submission Approval Letter, up to 1 page each, on headed paper, from the head of department or deputy / equivalent at the institution/s where the PI and any Co-Is (or other team members) are based, that includes;

  • support for the proposal
  • confirmation of the requested costs  
  • confirmation that the university will be willing to second the team members to Turing, under the terms and conditions outlined in the Secondment Agreement - see bottom of page to download a copy.

The Principal Investigator must ensure that a Submission Approval Letter is submitted for all universities on multi-party applications.  There is suggested text in the Submission Approval Letter template - see bottom of page to download a copy. 

4. Budget costings (see bottom of page to download spreadsheet) 

Applications

Applications must be submitted via the Flexigrant online portal

Please click on “Probabilistic Programming call". This should take you to the proposal form plus costings spreadsheet and Submission Approval Letter template. Costings and Approval Letters should be provided in the template format.

If you have not already done so, all applicants must first register on the system and provide basic details to create a profile. If you have any questions regarding the application form or using the online system, please contact Chiara Sbruazzo [email protected]    

We must receive your application before 13.00 on Tuesday 10 Oct  2023.

Documents available at the bottom of the page: 

  • Probabilistic Programming Call Submission Guidelines
  • Costings spreadsheet 
  • Submission Approval Letter template  
  • Sample Secondment Agreement  
  • Full Flexigrant application form

If you are employed at a university that is part of the Turing University Network, please contact your Turing liaison to make them aware of your application.

Assessment and review

The assessment and review process will be undertaken during the following stages: 

  1. Stage 1.0: Eligibility and triage 
  2. Stage 2.0: Expert review 

Following eligibility checks, proposals will be reviewed by an assessment panel who will rank the proposals based on the scores. 

The assessment panel will consider the following criteria: 

  1. Scientific Novelty, Timeliness and fit to call  
  2. Research team’s engagement with the Turing   
  3. Impact  
  4. Feasibility and value for money  

Each of the criteria will be scored and while all criteria will have equal weighting in evaluation, there will be a minimum requirement on significance to be considered for approval.  

Queries 

Please contact Chiara Sbruazzo, The Alan Turing Institute, Research Project Manager, via  [email protected]

Please monitor this page for updates.