What is the Discovery Engine?
It's a completely automated, end-to-end system for finding hidden patterns in data.
It allows us to find far more interesting and useful insights than is possible with existing data analysis methods: non-linear relationships between data of different modalities; complex patterns based on multiple variables; meaningful novel features and more.
This enables a new kind of scientific enquiry. Instead of formulating hypotheses and then gathering data to confirm or falsify them, we can begin with massive data and systematically extract all meaningful patterns from it. This not only minimises bias, subjectivity and irreproducibility, but is also much faster and can lead to more surprising discoveries.
Your data goes in, and insights come out.
Featured Articles
A New Scientific Paradigm
Science is Broken. Formulating hypotheses and designing experiments to confirm or falsify them is too slow, too unreliable, and far too path-dependent. Despite ever-increasing investment, progress is slowing. Most papers are wrong. Ideas are getting harder to find.
Discovering Hidden Patterns in Plant Biology
We recently partnered with the Montpellier Institute of Plant Sciences to apply the Discovery Engine to a tough, high-dimensional problem in plant biology: how do genotype and environmental conditions shape early root architecture in Arabidopsis thaliana?
Surfacing the Unexpected: Improvements in Meteorological Modelling
In a recent collaboration with the National Center for Atmospheric Research (NCAR)—a leading institution sponsored by the U.S. National Science Foundation—we applied the Discovery Engine to one of the most complex and dynamic systems on Earth: the coupled ocean-atmosphere boundary layer.
What makes the Discovery Engine different?
The standard scientific method is hypothesis-driven, laborious, and limited by human assumptions. Typical ML workflows provide opaque, task-specific automation without real understanding of the underlying patterns. Our Discovery Engine outstrips both, and lets your data drive the process, leveraging deep learning for uniquely superhuman pattern recognition ability. This gives you access to insights that existing methods would never find, and is completely automated - saving hundreds of hours of modelling and analysis time.
How does it work?
Deep neural networks often achieve beyond expert ability on predictive tasks, encoding complex, non-linear relationships and extracting meaningful features without human intervention. The problem is, they've always been considered black boxes - they find novel patterns that enable superhuman performance, but we can't then understand what they've learned. Our research changes that. Our automated system for training models optimised for knowledge discovery, combined with our novel interpretability algorithms, allow us to uniquely apply the pattern recognition ability of deep learning to systematically extract insights from complex data.
Join the waitlist!
We're inviting select researchers in academia and industry to use the Discovery Engine as part of our pilot study programme. If you would like to be considered for participation, please join the waitlist below.
FAQ
How mature is this technology?
We are at TRL6 and we have validated our core system with internal and external case studies. We are now beginning larger pilot studies with external partners.
Will this work for my domain?
The Discovery Engine is domain and modality agnostic, and can be adapted for most use-cases.
Do I need to understand neural networks?
Nope, that's our job. All you need is data.
Is this ChatGPT?
No. Language models are built for parsing text, not big complex datasets - we use proprietary deep learning and interpretability algorithms.