陶哲轩推荐AlphaEvolve，称其为数学发现的新工具。该系统解决了67个数学问题，超越人类最优解，自主发现新结构，展现出优越的可扩展性和鲁棒性，推动数学研究进展。

谷歌的AlphaEvolve打破了56年来的矩阵乘法效率记录，改进了Strassen算法，将4×4矩阵乘法的标量乘法次数从49次减少到48次。这项技术不仅能解决复杂数学问题，还能提升AI训练和芯片设计的效率。研究人员通过进化算法让AI自主探索，发现更优解，展现出自我优化的潜力。

谷歌DeepMind推出的AlphaEvolve是一款编程AI代理，结合大语言模型与自动评估，能够设计和优化算法，提升数据中心和芯片设计效率，发现新算法并解决数学问题。AlphaEvolve在多个领域展现出巨大潜力，未来将与研究团队合作开发用户界面，推动更多应用。

机器之心数据服务现已上线，提供高效稳定的数据获取服务，帮助用户轻松获取所需数据。

AlphaEvolve是一个基于大型语言模型的进化编码代理，旨在发现和优化算法。它通过创意和自动评估提升谷歌数据中心的效率，设计了更快的矩阵乘法算法，并在数学和计算领域取得显著进展。AlphaEvolve能够处理复杂问题，推动算法发现，未来有望在多个领域带来变革。

New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators

New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators

New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators

New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators

New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators

New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators

New AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluators