AI
Google DeepMind launches AI for Math Initiative with five leading research institutions
A new partnership aims to harness AI as a collaborative tool for mathematicians, accelerating discovery across theoretical and applied mathematics.
Google DeepMind Blog
Google DeepMind and Google.org have announced the AI for Math Initiative, a partnership with five world-leading research institutions to explore how artificial intelligence can accelerate mathematical discovery.
The inaugural partners are:
- Imperial College London
- Institute for Advanced Study
- Institut des Hautes Études Scientifiques (IHES)
- Simons Institute for the Theory of Computing (UC Berkeley)
- Tata Institute of Fundamental Research (TIFR)
A collaborative approach to mathematical progress#
The initiative frames AI not as a replacement for human mathematicians, but as an augmentation tool. Partners will work to identify mathematical problems suited to AI-driven insights, build supporting infrastructure, and accelerate the pace of discovery.
Google's contribution includes funding and access to cutting-edge AI systems:
- Gemini Deep Think, an enhanced reasoning mode
- AlphaEvolve, an agent for algorithm discovery
- AlphaProof, a formal proof completion system
Recent breakthroughs demonstrate capability#
The timing reflects real progress. In 2024, DeepMind's AlphaGeometry and AlphaProof systems achieved silver-medal performance at the International Mathematical Olympiad. More recently, the latest Gemini model with Deep Think reached gold-medal level, perfectly solving five of six IMO problems.
AlphaEvolve has shown particular promise in applied discovery. When tested on over 50 open problems across mathematical analysis, geometry, combinatorics, and number theory, it improved 20% of the previously best-known solutions. Notably, it discovered a new matrix multiplication algorithm that breaks a 50-year record: multiplying 4×4 matrices using just 48 scalar multiplications, surpassing Strassen's 1969 algorithm.
Opening new research pathways#
By combining human mathematical intuition with AI's novel capabilities, the initiative aims to open unexplored research directions and advance fundamental knowledge across science.
