The sustainable material development specialists, Orbital materials, are leveraging AI to design and deploy advanced materials and climate technologies. In their significant progress towards urban decarbonization, as the world grapples with an urgent need for sustainable solutions, they have released Orb which is an open-source AI model offering an innovative approach to simulating advanced materials.
It builds upon the company’s already existing foundational model ‘LINUS’, which has been developed entirely from scratch. This launch serves as a significant element in the company’s strategy to aid global decarbonization and leverage artificial intelligence to help scientists in creating the next generation materials which are essential for the energy transition. It allows them to peer into the interaction of materials at atomic level which the traditional microscopy falls short of, thus providing a new dimension for getting insights that are crucial for advancing the sustainable technology.
The CEO of Orbital, Jonathan Godwin, also highlighted that this ability to understand the mechanisms behind the properties of advanced materials will allow computers to design more effective materials. Simulating quantum physics is not easy, therefore, the present simulations only present a considerable simplification of what occurs at an atomic level. Nevertheless, he claims that Orb is much more accurate than the existing models by Google and Microsoft as it is able to achieve five times faster performance.
The introduction of the model Orb, not only highlights that Orbital is taking a lead in this field but also reinforces the significance of leveraging artificial intelligence to tackle such pressing environmental challenges.This model is being released by Orbital under a permissive open-source licence, so it is available for non-commercial uses and startups for free with the aim that this would impact and accelerate the sustainable efforts across global teams. Accompanying resources, including a supporting technical blog has also been provided and a technical report is expected to be published soon to to aid this model’s applications.furthermore, the technical description has been provided on the GitHub repository.
