Summary

  • Miles Cranmer wants artificial intelligence (AI) to help make scientific breakthroughs and speed up the current process.
  • Cranmer, who combines AI with his doctoral research in astrophysics at Princeton University, said the machine learning models lack “foundation models” that are designed for general scientific discovery.
  • To address this, Cranmer and more than 20 other scientists launched the Polymathic AI initiative at the start of this year to create foundation models equipped with scientific skills.
  • He said the advantage of these models is that they can generalise from new information and neural networks are unable to do this.
  • To address the lack of scientific data available for training, the team is turning scientific information into a format that can be used for training algorithms.
  • The next step will be to ensure results are explainable and of a high enough standard to be useful.

By John Pavlus

Original Article