Do we live in a simulation?
1st December 2022
Oxford, UK, 30 November
Some think we all live in a grand simulation. Certainly, computer simulations, Artificial Intelligence, and scenario modelling are having an increasingly large impact on our lives. Together, they are expanding and accelerating our knowledge and understanding of almost every aspect of science, research, life, and the World, even the Universe we inhabit.
At the forefront of this new era, with innovations in Neural Network technology, is Machine Discovery; a spin-out by Oxford University Innovation (OUI) the technology transfer office of the University. Machine Discovery is a rising star which is helping to tackle some of the big issues of our times. Their Discovery Platform is already proving instrumental in accelerating developments in fusion energy; the technology which harnesses the power of the Sun and promises limitless cheap energy. The platform is also being used to great effect in atmospheric climate modelling and by early adopters across diverse industries including semiconductor design.
Arising from foundation technology developed at the Department of Physics, University of Oxford, Machine Discovery was set up in 2019 by OUI
Co-Founded, by Professor Gianluca Gregori, Professor Sam Vinko, Dr. Muhammad Kasim, and Dr. Brett Larder, Machine Discovery’s Neural Network technology provides an easy-to-use software solution that interacts with virtually any simulator and modelling technology, intelligently learning, and optimising the simulation tasks. Discovery Platform then applies this learned, optimised model for future predictions, achieving high accuracy results with significantly faster run times. The initial algorithms developed by Machine Discovery garnered worldwide attention and coverage in high-profile academic journals. The Company has built on this technology and is now targeting large commercial contracts, initially in the fast-growing Electronic Design Automation (EDA), Engineering, and Clean Energy markets.
In stark contrast to ‘traditional’ AI and deep learning algorithms, the Discovery Platform offers significantly faster ‘learning’. Where tens of thousands of learning data are typical, Discovery Platform can achieve virtually identical results from significantly fewer training data points and in some cases, such as simulations in Global Aerosol-Climate Modelling, simulations were accelerated by >1 bn times. More typical accelerations are seen in the 10,000 to 10 million times range, leading to reduced project times and faster time to market.
Bijan Kiani, CEO of Machine Discovery commented: “Our open, easy-to-use AI-powered platform is being deployed to augment existing simulators for prediction and optimisation of the toughest challenges in research. Already, we’re seeing fusion projects reduced from twelve-month duration to only two months. Machine Discovery are delivering the next step change in productivity”
Editorial Contacts:
Oxford University Innovation
Machine Discovery
+1 650-283-6917
SOURCE Oxford University Innovation
Press release sign up
