latentbrief
Back to news
Launch1d ago

AI Tool Speeds Up Scientific Research By Writing Code For Scientists

Google AI Research, Digg AI1 min brief

In brief

  • A new artificial intelligence tool called ERA, developed by Google, is revolutionizing scientific research by writing and optimizing scientific code.
    • This tool, which uses Gemini, helps scientists save time by automatically generating and refining computational experiments.
    • It was tested across various fields like genomics and neuroscience and achieved expert-level performance in all areas.
  • ERA works by taking a problem and a success measure, then exploring thousands of solutions using a tree search approach to optimize the code.
  • Google researchers have already used ERA on eight projects, including epidemiological forecasting and satellite imagery analysis, demonstrating its potential to democratize access to expert-level computational modeling.
  • Scientists can now apply ERA through a trusted tester program in Google Labs.
    • This tool could significantly accelerate scientific discovery by making complex coding more accessible.
  • Watch for ERA's broader availability and its impact on various scientific fields as it continues to evolve.

Terms in this brief

ERA
ERA is a new AI tool developed by Google that writes and optimizes scientific code to speed up research. It uses Gemini to generate and refine computational experiments, achieving expert-level performance in fields like genomics and neuroscience.
Tree search approach
A method where ERA explores thousands of potential solutions to find the best one for a given problem and success measure. This systematic exploration helps optimize scientific code effectively.

Read full story at Google AI Research, Digg AI

More briefs