Google DeepMind’s AI Agent has unveiled a groundbreaking artificial intelligence (AI) agent capable of developing algorithms that surpass even the best human-designed solutions. This development represents a significant milestone in computational science, where machines not only follow instructions but now actually create them—optimizing and solving tasks once thought to require human intuition and decades of research.
Image Credit : Getty Images
Google DeepMind shows that with a few clever tweaks these models can at least surpass human expertise designing certain types of algorithms—including ones that are useful for advancing AI itself.
How Google Uses Large Language Models to Discover New Algorithms
The company’s latest AI project, called AlphaEvolve, combines the coding skills of its Gemini AI model with a method for testing the effectiveness of new algorithms and an evolutionary method for producing new designs.
AlphaEvolve came up with more efficient algorithms for several kinds of computation, including a method for calculations involving matrices that betters an approach called the Strassen algorithm that has been relied upon for 56 years. The new approach improves the computational efficiency by reducing the number of calculations required to produce a result.
Traditionally used for natural language processing tasks like translation and summarization, large language models are now being repurposed for algorithmic discovery. By training these models on vast datasets of mathematical problems and solutions, DeepMind enables its AI agent to understand the structure of problems and propose optimized solutions.
