A synergy of artificial intelligence (AI) and robotics, embodied in Google (NASDAQ:GOOGL) DeepMind's A-Lab, has made significant strides in material science, autonomously synthesising new materials with potential uses in energy and technology sectors.
This groundbreaking AI system from Google DeepMind has predicted nearly 400,000 stable materials from 2.2 million potential compounds, offering an expansive horizon for future material creation.
The system dubbed the "ChatGPT for materials discovery,” combines the predictive power of AI with the precision of robotics, potentially revolutionising the material discovery process.
Notably, the A-Lab operates by autonomously devising recipes for materials and then synthesising and analysing them, a process entirely free from human intervention.
Introducing GNoME: an AI tool that helped discover 2.2 million new crystals. ????Crystals are found in everything from the chips powering our phones to solar cells creating clean energy.
The model also better predicts the stability of new materials. ???? https://t.co/O3YdnVcJt1 pic.twitter.com/OZZjZxf9dd
— Google DeepMind (@GoogleDeepMind) November 29, 2023
Transformative impact
Ekin Dogus Cubuk, who leads the materials discovery team at Google DeepMind in London and is involved in both studies said: “A lot of the technologies around us, including batteries and solar cells, could really improve with better materials.
"This statement underscores the transformative impact these new materials could have on various technologies."
381,000 new inorganic compounds
Traditionally, the discovery of new materials has been a slow and meticulous process.
However, Google DeepMind's AI system, named GNoME (graph networks for materials exploration), represents a paradigm shift.
By analysing and modifying known materials, GNoME has proposed over 2.2 million potential compounds.
Following assessments of stability and crystal structures, this figure was refined to a staggering 381,000 new inorganic compounds.
About the A-lab
Situated at the Lawrence Berkeley National Laboratory (LBNL), the A-Lab showcases the pinnacle of robotic technology in material synthesis.
The A-Lab uses an extensive database of over 30,000 synthesis procedures, enabling it to recommend precise ingredients and reaction temperatures.
An 'active learning' algorithm further refines these processes, allowing for the iterative and successful creation of new materials.
In a 17-day span, the A-Lab produced 41 new inorganic materials, 9 of which were synthesised following improvements by the active learning algorithm.