From ideas to fundable AI projects: workshop recap and next steps
Jawwad Arshad Darr, Emma Tobin, and Matt Davis opened the session by introducing the current funding scene and how interdisciplinary teams can compete.
The room brought together researchers from across MAPS with one goal: turn early concepts into credible, fundable projects.
Who was in the room
- Chemistry and Materials Discovery
- Mathematics and Statistical Science
- Physics and Astronomy
- Science and Technology Studies
- Earth Sciences
- Risk & Disaster Reduction
This mix made it easy to test ideas, spot shared challenges, and form new collaborations on the fly.
Themes that gained traction
- AI-enabled materials discovery
- Predictive modelling and uncertainty quantification
- Data-driven methods for experimental design
- Ethics and responsible innovation
Before the networking session, participants exchanged quick-fire pitches to surface overlaps and gaps. That sped up team formation and focus.
What moved the needle
- Structured activities that pushed ideas from "interesting" to "testable."
- Clear problem framing tied to measurable outcomes.
- Early attention to data availability, baselines, and feasibility.
- Embedding responsible innovation from the start, not as an afterthought.
Outcomes
Through these sessions, at least four consortia are now moving ahead to develop their strongest ideas into grant submissions over the coming months.
By the end of the workshop, several early-stage proposals had real shape-evidence of strong appetite and potential for cross-disciplinary AI research at UCL.
The MAPS Enterprise team will continue supporting these groups as proposals progress. More workshops are planned this year to develop grants on additional priority topics.
If you plan to join the next sessions, prep this
- Two short problem statements (what's the scientific gain and why AI now?).
- A quick view of data: access, quality, size, and gaps.
- Baseline or heuristic you aim to beat and how you'll measure success.
- Potential collaborators across methods, domain science, and impact users.
- A one-paragraph note on ethics and responsible innovation (risks, mitigations, governance).
Useful resources for research teams
- AI for Science & Research - examples, tools, and methods relevant to materials, modelling, and data-driven science.
- AI Learning Path for Research Scientists - practical steps for modelling, lab automation, and experiment design.
- Machine learning for materials discovery (review) - a concise overview of methods and use cases.
Bottom line
Diverse teams, clear problem statements, and early attention to data and ethics turned ideas into credible proposals. With several consortia now moving to submission, the signal is clear: cross-disciplinary AI research here has momentum-and support to match.
Your membership also unlocks:
