Unverified50% confidenceFactExact time
RAG (Retrieval-Augmented Generation) was first proposed by a Meta research team in 2020; its core idea is to introduce external knowledge retrieval before large model inference to mitigate hallucination issues.
1
Sources
50%
Confidence
Long-term
Relevance
9/10/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRAG addresses the two major pain points of pure generative models: outdated knowledge and factual hallucination.67% similarUnverifiedAround 2020, 'second brain' tools represented by Roam Research and Obsidian rose to prominence, shifting the core logic toward 'connection and emergence.'63% similarUnverified闭环优化这一思路最早应用于脑机接口领域，后逐渐渗透到感知神经科学研究中62% similarUnverifiedThe theoretical basis of knowledge distillation comes from the groundbreaking work of Hinton et al. in 2015, which transfers dark knowledge contained in soft labels to smaller models.59% similarUnverifiedThe idea of expectation-based localization (integral regression) was first proposed by Jian Sun's team at SenseTime in the task of human pose estimation.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/891619API
curl https://kongchang.com/api/v1/knowledge/claims/891619MCP
get_claim(id=891619)