A group of international scientists has joined forces to unveil a groundbreaking research initiative called Polymathic AI, harnessing the technology that underpins ChatGPT to create an artificial intelligence tool tailored for scientific discovery.
ChatGPT predominantly deals with natural language processing, but Polymathic AI aims to excel in a different realm. It will be built upon numerical data and physics simulations spanning various scientific disciplines, supporting researchers in modelling diverse subjects such as supergiant stars and climate patterns.
Principal investigator Shirley Ho, who is a group leader at the Flatiron Institute’s Center for Computational Astrophysics in New York City, believes this initiative will revolutionize the application of AI and machine learning in the field of science. She likens the concept to the ease of learning a new language when already familiar with several others.
Polymathic AI capitalizes on a pre-trained model, known as a foundation model, which can significantly expedite and enhance the development of scientific models compared to starting from scratch. Even when the training data might not seem directly relevant to the problem, this approach can unveil commonalities and connections between various scientific domains.
The Polymathic AI team comprises experts from physics, astrophysics, mathematics, artificial intelligence, and neuroscience. Their project will collect data from a multitude of sources across physics and astrophysics, with the intention to expand to fields like chemistry and genomics. This multidisciplinary approach will be applied to a wide range of scientific challenges.
ALSO READ: Is your smartphone in WhatsApp’s list of devices losing support after October?
One significant improvement Polymathic AI offers over ChatGPT is its focus on precision. The project treats numerical data as genuine numbers, not merely characters, and employs real scientific datasets that encapsulate the underlying physics of the universe.
Transparency and openness are central to Polymathic AI’s mission. The team aims to make all their work publicly accessible, with the goal of democratizing AI for scientific research. Ultimately, they plan to offer a pre-trained model to the scientific community, which can enhance scientific analyses across diverse problem areas and domains. This project represents a significant step toward broadening the application of AI in the realm of scientific discovery.
ALSO READ: Microsoft encourages researchers to find vulnerabilities in AI-powered Bing
Inputs from IANS
