Unverified50% confidenceFactExact time
Orthogonal polynomials are typically defined on bounded or semi-infinite intervals, whereas real inputs to neural networks are often unbounded real values, creating a gap between the two.
1
Sources
50%
Confidence
Long-term
Relevance
10/4/2026
First Seen
Sources
Related Entities
Related Claims
Unverified线性代数中的矩阵分解与特征值分解对应神经网络的权重空间分析72% similarUnverifiedAfter neuron alignment, the loss landscape between two networks often exhibits a linear pattern, allowing interpolated models to maintain considerable performance, whereas naive averaging typically falls into high-loss barrier regions.69% similarUnverifiedIn the unfused small-operator mode, Cube computation and Vector computation are often arranged serially, causing one type of compute unit to be idle at any given moment.68% similarUnverified梯度学习的核心是连续优化，神经网络参数空间是高维实数空间，损失函数近似可微67% similarUnverifiedNaively approximating the infinite-dimensional functional gradient causes the algorithm to converge to the wrong location, introducing systematic bias.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/971800API
curl https://kongchang.com/api/v1/knowledge/claims/971800MCP
get_claim(id=971800)