top of page
SERVICES
Trustworthy AI
Structure is everywhere to be discovered in both the natural world and human society. The recognition and transformation of structural patterns underpins human intelligence and is the essential machinery of both scientific understanding and robust solutions to real-world problems.
Hylomorph’s unique approach to trustworthy and explainable Artifical Intelligence
places structure at the core of the learning process.
Our learning algorithms address key technical and societal concerns of
causality, generalizability, fairness and safety.
CLIENTS
How to make Machine Learning speak the language of your business, generalize better and learn faster?
Our clients and collaborations
In collaboration with Level 42 AI, we applied our sample-efficient learning algorithms to achieve state-of-the-art accuracy at detection of pulmonary diseases in five major clinical studies carried out on three continents.
We help Kardiolytics take medical imaging to the next level by leveraging state-of-the-art ML and combining it with our structure-aware approaches.
We help So Energy, a London-based energy supplier focusing on renewables, to develop intelligent, adaptive and proactive customer-facing services. Our solution provides principled analysis of customer feedback expressed in natural language, preventing misinterpretation, hallucination and bias by design.
Hylomorph is an industrial partner in EPSRC's £80m investment in next-gen AI, supporting research into mathematical and computational foundations:
https://www.ox.ac.uk/news/2024-02-06-new-oxford-research-hub-propel-transformative-ai-innovations
We collaborate with the Department of Mathematical Sciences at Durham University, working on algebraic dynamical systems in Machine Learning. See the recent publication summarising the results:
https://link.springer.com/article/10.1007/s10485-023-09762-9
Recent developments
The head of our business development gave a talk on "work-on-command" at the Davos Innovation Week, a satellite event of the World Economic Forum (WEF'2024).
We can help you if:
You need a software solution to a problem which our experts can describe in terms of classification, regression, optimization or decision support.
This includes all problems which are addressed by conventional machine learning, and many more besides.
Need more details?
bottom of page