The traditional process of drug discovery is too long, too expensive and too risky.
We believe AI can improve it – helping to get better drugs to the people who need them, and to treat and cure disease faster.
If we think of biology as fundamentally an information processing system – one that transmits information and maintains structure – we can start to see how it might share a basic underlying structure, or an ‘isomorphic mapping’, to information science.
If this conjecture proves true, we should be able to apply the rapid pace of machine learning progress to the modelling of the principles of biology – with extraordinary, and extraordinarily fast, results.
At Iso, we are building on Google DeepMind’s pioneering research, and work in partnership with our colleagues there as members of the Alphabet family. There, they used games like StarCraft and Go as a testing and learning ground for AI capabilities, and showed the impact it can have on many scientific fields – from structural biology and quantum chemistry to fusion and mathematics.
Thanks to their groundbreaking work – and their revolutionary systems like AlphaFold, a solution to the 50-year grand challenge of protein folding – we now know that artificial intelligence can be used for more than just analysing data. It can be reliably used to build predictive and generative models of complex biological phenomena. It can learn the rules of biology.
And that’s what we’re doing here: developing cutting-edge computational techniques in fields like deep learning, reinforcement learning, active learning, representation learning and more to solve some of the toughest challenges in drug discovery, and some of the most stubborn scientific problems in biology, chemistry and medical research today.
The result – we’re pretty sure – will be to deepen our understanding of so much of the world and ourselves; of what ails some of us and what might cure more of us.
We’re on a path to unlock the full potential of machine learning for drug discovery – and beyond.
This is the promise of digital biology at Iso.
The success of this project relies on our brilliant teams working together, learning from and empowering each other every single day.