One scientific system · Three distinct contributions

From limited resources to reliable scientific knowledge

How can limited observations, experiments, and computational resources become reliable, testable scientific knowledge that future research can build on?

We seek to understand and improve how scientific discovery happens. Our long-term mission is to develop principles and methods of scientific learning and discovery grounded in mathematics, information, and computation. We ask how scientific systems can choose meaningful variables and scales, distinguish competing explanations, acquire informative evidence, and revise models, concepts, and questions as new evidence emerges.

Our research connects the design of scientific learning systems, the creation and acquisition of evidence, and mathematical and data-science exploration.

Current focus: mechanistic questions in molecular and materials science

We develop AI methods to understand and discover mechanisms, and to use that understanding to guide design. Our current focus is how structure, interactions, and evolving processes give rise to properties, functions, and responses in molecular and materials science.

Feedback from experiments and simulation helps us test and improve these methods while advancing scientific understanding. As methods are repeatedly validated, we will extract broader mathematical, computational, and system-design principles and test where they transfer.

Three connected contributions

Distinct scientific questions. Shared discovery logic.

Each Pillar addresses a different part of scientific discovery. Together, they connect how systems learn, what evidence they need and which mathematical ideas can take them further.

Pillar I The scientific learner

Design how scientific learning systems work

Foundational Scientific Learning-System Design

We design scientific representations, model architectures and information flow, learning signals and objectives, and the ways models learn and adapt.

Explore Pillar I Try changing the connections
Research focus and contribution · Pillar I

How should a scientific system learn?

Research focus

  • Representation spaces that preserve relevant structure
  • Interaction structures that expose meaningful relations
  • Guiding signals aligned with scientific questions
  • Learning dynamics for stable and reusable knowledge
  • Scientific system architecture, research infrastructure, and evaluation
Intended contribution

A learner whose internal organisation can support scientific interpretation, testing, and design.

Pillar II The evidence environment

Design what evidence systems acquire

Evidence Engineering and Closed-Loop Discovery & Design

We design datasets, perturbations, candidates, and experiments that distinguish competing explanations and change what can be learned next.

Explore Pillar II Try choosing the next experiment
Research focus and contribution · Pillar II

What evidence should the system create next?

Research focus

  • Data quality, provenance, and decision cohorts
  • Data factories and evidence-producing workflows
  • Perturbation and experimental design
  • Active acquisition and information gain
  • De novo, inverse, and closed-loop design
Intended contribution

An evidence strategy that reduces uncertainty and guides the next scientific decision.

Pillar III The mathematical lens

Explore mathematics and data science for scientific discovery

Mathematical Data Science and Frontier Exploration

We investigate mathematical and data-science questions and develop methods that reveal structure in complex observations.

Explore Pillar III Explore what the observations identify
Research focus and contribution · Pillar III

What mathematical and data-science methods can reveal important structure and open new scientific possibilities?

Research focus

  • Mathematical modelling
  • Statistical inference
  • Scientific computing
  • Analysis of complex data
  • Assumptions, computational limits, and transfer
Intended contribution

Principles and tools that reveal when patterns can become testable explanations or useful designs.

Integrated research map

The Pillars are contributions to one system—not silos or stages.

Pillar III
Mathematical lens Mathematical modelling · scientific computing · data analysis
Pillar II Evidence environment Datasets · perturbations · experiments · candidates
Pillar I Scientific learner Representations · models · systems · evaluation
Scientific contribution Mechanism & design Explanation · control · artifact · next question
Scientific testbeds Molecular interactions and drug discovery Proteins, peptides, and sequences Biomaterials and delivery systems Complex scientific data and mechanisms
Pillar I designs scientific learning systems, their infrastructure, and evaluation. Pillar II creates and acquires evidence. Pillar III develops mathematical and data-science methods. Scientific questions connect all three contributions.
View the conceptual illustration
Three connected groups show a blue network across representation planes, observation tiles beside a revisable model, and a curved mathematical surface intersected by constraint planes.

A conceptual view of learning-system design, evidence engineering, and mathematical exploration as three connected contributions to scientific discovery.

Scientific learner
Design representations, models, systems, and scientific evaluation.
Evidence environment
Design what to measure, test, and learn from next.
Mathematical lens
Reveal structure, uncertainty, and new scientific questions.

Research horizon

Build now. Connect next. Generalise carefully.

These are planning windows. Expansion will follow research progress and validation.

  1. Now

    Mechanistic questions in molecular and materials science

    We develop and test AI methods for mechanism understanding, discovery, and design through scientific questions and feedback from experiments and simulation.

  2. Over approximately five years

    Building reusable scientific capabilities

    We aim to develop reusable representation and model methods, scientific system architectures, research infrastructure, evidence-acquisition strategies, and evaluation systems that support mechanism understanding and reliable design.

  3. Over approximately five to ten years

    Testing principles across systems

    We aim to extract mathematical, computational, and system-design principles and test how these methods and capabilities transfer and combine across broader scientific fields.

Shared scientific principles

Principles for reliable discovery.

These principles connect the questions we ask, the methods we develop and the evidence that shapes our understanding.

01

Mechanism-aligned

Representations and objectives should preserve distinctions that matter to the scientific mechanism.

02

Evidence-efficient

The next dataset or experiment should reduce uncertainty that matters to a decision.

03

Uncertainty-aware

A scientific system should expose what it does not know and where its conclusions may fail.

04

Compositional

Reusable primitives, interactions, and operators should support analysis across systems and scales.

05

Testable

Explanations should imply controls, perturbations, or observations that could challenge them.

06

Design-oriented

Learned structure should return to experiment through candidate, molecule, material, or system design.

Continue through the evidence

Follow the programme through outputs, people, and collaboration.

Read about our approach

OutputsExplore publicationsDiscover the work behind our research. PeopleMeet the researchersFind the people developing each part of the programme. CollaborateBring a testable questionStart a scientific, experimental, or recruitment conversation.