Unverified50% confidenceFactExact time
The splines on the edges of a KAN are only locally effective within their domain, and this locality makes KAN's changes to existing functions more concentrated when learning a new task, so it is less likely to affect the parameter regions associated with old tasks.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn token-level sequence generation tasks, the KL constraint is computed token by token, so tokens at different positions are subject to different constraint strengths.72% similarUnverifiedKAN (Kolmogorov-Arnold Networks) places learnable univariate functions on the edges of the network, rather than placing fixed activation functions on the nodes as in traditional neural networks.66% similarUnverifiedIn EDMD-kDL, the outer optimization learns the parameters of the kernel dictionary, and the inner optimization solves for a finite-dimensional approximation of the Koopman operator under that dictionary, with the two iterating alternately.66% similarUnverifiedEdit Locality measures a model's ability to keep unaffected regions unchanged when executing local modification instructions; many diffusion models cause subtle subject distortion or semantic drift due to global resampling.66% similarUnverifiedKAN places learnable activation functions on the edges (connection weights) of the network, typically implemented as B-spline functions, whereas traditional MLPs use fixed activation functions at the nodes.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1073723API
curl https://kongchang.com/api/v1/knowledge/claims/1073723MCP
get_claim(id=1073723)