Anthropic 与哈佛用 Claude Mythos 5 优化量子电路设计
Anthropic 和哈佛用 Claude Mythos 5 自动搜量子电路，8 个任务里 7 个超过了人类最佳设计，每个结果都经过独立验证，不是嘴上说说。
Anthropic、哈佛等机构联合发表研究，用 OpenEvolve 搜索电路生成程序，由 Claude Mythos 5 提出量子电路架构与程序改写方案。LLM 提案、独立评估器验证正确性并计算成本，最优设计进入下一轮迭代。在 8 项任务中的 7 项上，LLM 提出的电路超过了已知最佳设计，使用的 Toffoli 门和量子比特更少。任务覆盖数值计算、化学、优化、量子模拟和数论。在第 8 项任务上，搜索发现了素数域模逆运算的新可逆架构，让经典算法 divsteps 可用于可逆算术并改善了 Toffoli-qubit 权衡。
Banger paper from Anthropic, Harvard, and colleagues.
They find that LLM-proposed quantum circuits beat the best-known designs on seven of eight tasks, with fewer Toffoli gates and fewer qubits.
They search over these circuit-generating programs with OpenEvolve, using Claude Mythos 5 to propose architectures and program rewrites.
LLMs propose circuit designs, an independent evaluator verifies correctness and scores the cost, and the best designs seed the next round.
The tasks cover numerical computation, chemistry, optimization, quantum simulation, and number theory.
On the eighth task, the search found a new reversible architecture for prime-field modular inversion. It makes divsteps, a classical algorithm impractical for reversible arithmetic, usable and improves the Toffoli-qubit trade-off.
The evaluator verifies every circuit before scoring, so each reported gain is a checked circuit.
https://t.co/8hHTwMfKHS
Chat with Paper: https://t.co/z5TxCnpauU
