Contact: Dr. Gareth Conduit, Physics & Ben Pellgrini, Intellegens
Mentor: Paul May
Intellegens is a spin-out from the University of Cambridge which develops unique Artificial Intelligence (AI) algorithms that have the capability to train and predict from incomplete data. The algorithm has been developed in the Physics department and has already had commercial success in the drug discovery and material design markets. In both these domains long product development cycles require many experiments and trials.
Intellegens are able to cut costs by making predictions based on incomplete experimental data. This improved data reduces the number of experiments needed, leading to improved efficiency and accelerating the product to market.
While the impact of the algorithms is already demonstrable in the domains of drug discovery and materials design, the model and approach is generic and could be applied to any sectors where there is big, sparse data, such as finance, retail, Internet of Things (IoT), weather, military, security, and many others.
Intellegens are currently developing a service to allow for the wider commercialisation of this technology, utilising modern cloud and web based technologies. This would allow customers to run their own experiments and integrate the software into their own software / modelling pipeline systems.
The challenge for the i-Team is to investigate and analyse the wide range of potential uses for these new AI algorithms, looking for any application that needs to make use of incomplete datasets and make predictions from them.