Google researchers use AI and Lean proofs to reduce quantum-program resource costs
Researchers at Google DeepMind and Google Quantum AI describe an AI-assisted method for designing quantum programs whose correctness and resource costs are checked with the Lean theorem prover. The preprint first appeared November 6 and was revised November 9. Read the paper.
The workflow starts with researchers fixing a mathematical specification. AI agents propose programs and proofs, share results, and submit candidates to a separate verifier intended to prevent them from weakening the task.
The authors report resource reductions of roughly 11 times for a fusion-related simulation, eight times for photochemistry and 2.5 times for drug-metabolism calculations. For fixed-point arithmetic, they report a reduction exceeding 290 times under their qubit-times-T-gate cost measure. Research PDF.
These are reported mathematical resource improvements, not measured speedups on commercially available quantum machines. A proof establishes compliance with the chosen specification; its relevance to a physical problem still depends on that specification. The research is a preprint, and Guth has not independently reproduced its results.
Sources
Sources and citations
Revision history
Published .