By Thang Luong and Vahab Mirrokni
Publication Date: 2026-02-11 15:00:00
Working with experts on 18 research problems, an expanded version of Gemini Deep Think helped solve long-standing bottlenecks in algorithms, ML and combinatorial optimization, information theory, and economics. Highlights from our article “Accelerating Research with Gemini” include (corresponding section numbers in the article):
- Pushing mathematical boundaries for network puzzles: Progress on classic computer science problems such as “Max-Cut” (efficient partitioning of networks) and the “Steiner Tree” (connecting high-dimensional points) had slowed. Gemini both broke through blocks by thinking outside the box. It solved these discrete algorithmic puzzles by drawing on advanced tools—such as the Kirszbraun theorem, measure theory, and the Stone-Weierstrass theorem—from entirely unrelated branches of continuous mathematics. See sections 4.1 and 4.2.
- Solving a decades-old conjecture in submodular online optimization: A 2015 theory paper proposed a seemingly obvious rule for data…