Unverified50% confidenceFactExact time
FunSearch replaces the random mutation operator in traditional genetic algorithms with a large language model, giving the mutation operation semantic-awareness capability.
1
Sources
50%
Confidence
Long-term
Relevance
8/10/2026
First Seen
Sources
AI攻克25年数学难题：人工智能如何改变数学研究
Related Claims
UnverifiedAlphaEvolve的创新在于用大语言模型替代随机变异，生成有意义的变异64% similarUnverifiedA Reddit developer used a genetic algorithm combined with a multilayer perceptron (MLP) to optimize the boarding problem, achieving a result 9.6% faster than the Steffen method in simulation.63% similarUnverifiedAlphaEvolve结合大语言模型与进化算法，让LLM生成候选代码，通过自动化评估器打分并迭代择优63% similarUnverified模型的平台风格适配依赖世界知识泛化而非硬编码规则，因此存在推断偏差的可能，仍需人工复核62% similarUnverifiedLarge language models predict the next token one at a time through an autoregressive generation mechanism, and are naturally inclined to generate more complete rather than more concise code.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/725253API
curl https://kongchang.com/api/v1/knowledge/claims/725253MCP
get_claim(id=725253)