Unverified50% confidenceFactExact time
K-Means' within-cluster sum of squares J is monotonically non-increasing in each iteration; since J has a lower bound of 0, the algorithm necessarily converges to a local minimum, but global optimality is not guaranteed and the result depends on initialization.
1
Sources
50%
Confidence
Long-term
Relevance
11/8/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn a K-Means implementation, an empty cluster (a center point to which no samples are assigned) requires reinitializing that center, otherwise the algorithm will be interrupted by a division-by-zero or empty-array error.75% similarUnverifiedWhen conditional dependencies exist between positions, the KL divergence between the product distribution and the true joint distribution is strictly greater than zero, and it increases as dependencies grow stronger and more positions are written simultaneously.71% similarUnverifiedThe residual structure makes F=0 the default option—as long as the F(x) weights are close to zero, the residual block degenerates into an identity mapping, thereby resolving the optimization difficulty of SGD in learning identity mappings.71% similarUnverifiedThe Hellinger modulus integrates the difference of the square roots of the distribution densities, and compared with KL divergence, it is bounded and often yields tighter constant bounds in the derivation of minimax lower bounds for nonparametric estimation.68% similarUnverifiedThe choice of k=8 is a compromise between cost and statistical power: the larger k is, the narrower the confidence interval, but the number of generation calls also grows linearly.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1149341API
curl https://kongchang.com/api/v1/knowledge/claims/1149341MCP
get_claim(id=1149341)