Verified75% confidenceFactExact time
The revival of modern MoE can be traced back to the 2017 paper published by the Google Brain team, 'Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer,' which was the first to integrate sparse MoE layers into large sequence models.
3
Sources
75%
Confidence
Long-term
Relevance
7/23/2026
First Seen
Sources
Qwen3 235B A22B实测：2.4万亿参数开源模型超越Claude Opus
Related Claims
UnverifiedGoogle Brain于2017年在《Outrageously Large Neural Networks》论文中将MoE规模化应用于Transformer框架79% similarVerifiedThe Mixture of Experts (MoE) architecture was scaled up within the Transformer framework by Google Brain in 2017; its core idea is to group model parameters into multiple expert sub-networks, with only a small subset activated during each inference.78% similarUnverifiedThe Google Brain team published research on "Neural Cellular Automata" in 2020, using neural networks to learn local update rules so that cellular automata can self-organize to grow a target pattern and self-repair after damage.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/600088API
curl https://kongchang.com/api/v1/knowledge/claims/600088MCP
get_claim(id=600088)