Unverified50% confidenceFactExact time
混合专家（MoE）架构最早由 Jacobs 等人于 1991 年提出，其核心思想是将复杂模型拆分为多个专家子网络，由门控网络根据输入动态决定各专家权重组合
1
Sources
50%
Confidence
Long-term
Relevance
7/17/2026
First Seen
Sources
Related Claims
UnverifiedThe core idea of the Mixture of Experts (MoE) architecture originates from the divide-and-conquer learning theory of 1991.74% similarUnverifiedThe Mixture of Experts (MoE) model is a classic machine learning concept originating in 1991.72% similarUnverifiedMoE 架构源自1991年 Jacobs 等人的学术论文《Adaptive Mixtures of Local Experts》68% similarUnverified任务路由的思想根源可追溯到早期专家系统和集成学习中的模型选择与混合专家（MoE）机制62% similarVerifiedGAN (Generative Adversarial Network) was proposed by Ian Goodfellow et al. in 2014; its training framework consists of two mutually competing neural networks: a generator and a discriminator60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/544376API
curl https://kongchang.com/api/v1/knowledge/claims/544376MCP
get_claim(id=544376)