Unverified50% confidenceFactExact time
The concept of intrinsic dimension was systematically proposed by Li et al. in their 2018 research.
1
Sources
50%
Confidence
Long-term
Relevance
10/9/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedDeepMind published research in Nature in 2021 in which AI discovered the functional relationship between knot hyperbolic volume and algebraic invariants.62% similarUnverifiedScaling Laws were formally proposed in 2020 by OpenAI researcher Jared Kaplan and others, establishing that model performance follows a power-law relationship with parameter count, data volume, and compute.57% similarUnverified傅里叶神经算子（FNO）由Zongyi Li等人于2020年提出，学习的是无穷维函数空间之间的映射，能以比传统数值方法快数个数量级的速度求解偏微分方程57% similarUnverifiedYann LeCun proposed the JEPA framework, whose core idea is to perform prediction in an abstract representation space rather than pixel space, and Meta has released early implementations such as V-JEPA.56% similarUnverifiedThe theoretical foundations of diffusion models can be traced back to 2015 when Sohl-Dickstein et al. first introduced diffusion processes into deep learning, with a breakthrough in image generation achieved after Ho et al. proposed DDPM in 2020.54% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/991078API
curl https://kongchang.com/api/v1/knowledge/claims/991078MCP
get_claim(id=991078)