Unverified50% confidenceFactExact time
On the corresponding truncated network class, empirical risk minimization (ERM) achieves the minimax regression rate n^{-2s/(2s+D)} with no additional logarithmic factor.
1
Sources
50%
Confidence
Long-term
Relevance
10/19/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedUnder a margin condition with exponent β, using buffered ERM calibrated with a margin buffer achieves excess risk on the order of Q^{-1/2}+N^{-β/2} (ignoring logarithmic factors).78% similarUnverifiedWhen the label error rate is below about 10%-20%, deep networks can still learn effective representations, and the degradation in generalization performance is limited.63% similarUnverifiedThe least squares method solves for regression coefficients by minimizing the residual sum of squares (RSS), where residuals refer to the difference between actual values and predicted values62% similarUnverifiedCertain AI capabilities are nearly zero when model scale is below a specific threshold, then rise sharply once the threshold is crossed, exhibiting a nonlinear 'phase transition' characteristic.62% similarUnverifiedThe researchers proposed Minimax RR-ICM, which optimizes over the worst case over probability distributions on the sphere rather than averaging over directions.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1068310API
curl https://kongchang.com/api/v1/knowledge/claims/1068310MCP
get_claim(id=1068310)