Unverified50% confidenceFactExact time
Three methods are commonly used in scaling-law fitting practice—least squares with fixed E, fitting with E as a free parameter, and linear regression in log space—and the optimal ratios they yield may differ.
1
Sources
50%
Confidence
Long-term
Relevance
11/4/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedUnder the setting of a shared regression function and commutable power-law covariance spectra, there exists a region where the scaling law achieves the minimax-optimal exponent.65% similarUnverifiedSince the joint lower bound shows that the dependence on (d, δ, ε) cannot be optimized in a decomposed manner, the trade-offs in algorithm design should be placed on improving constants and practical sample reuse efficiency, rather than attempting to break through the product structure of the three in worst-case guarantees.64% similarUnverifiedIn the adversarial environment, the regret of the paper's algorithm is \tilde{\mathcal{O}}(δ_*^{1/3} T^{2/3}), where δ_* is the fractional weak domination number.64% similarUnverifiedUp to constant scaling, every admissible non-decreasing rate curve can be achieved in both the realizable and agnostic settings.62% similarUnverifiedFor heterogeneous asynchronous MMAB, a polynomial-time greedy algorithm can achieve a 1/e approximation ratio relative to the optimal centralized assignment.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1133880API
curl https://kongchang.com/api/v1/knowledge/claims/1133880MCP
get_claim(id=1133880)