Unverified50% confidenceFactExact time
The theoretical analysis gives upper bounds on the squared ℓ2 parameter estimation error: Õ(s*/n + s*^2/(n^2ε^2)) for Sample-DP and Õ(s*/n + s*/(nε^2)) for Label-DP (based on randomized response), where n is the amount of training data and ε is the privacy budget.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe paper proves that the learned generative distribution satisfies a finite-sample error bound in Wasserstein-p distance: W_p ≲ n^(-1/d) + n^(-1/(2p))(log(1/ξ))^(1/(2p)).74% similarUnverifiedUnder the assumption that the true mapping is uniformly random, the paper proves a lower bound on the error rate of E ≥ (M/N)δ*(B/M) + (1-M/N)(1-1/K).73% similarUnverifiedUnder the total variation metric, the average error is Õ_p(n^(−β/(d_x+d_y+2β))).72% similarUnverifiedThe paper studies two classes of estimators: the preliminary estimator and the constrained estimator; the preliminary estimator trains a separate binary regression model for each measurable set, which is simple to implement and highly parallelizable, but predictions across different sets may violate probabilistic consistency.71% similarUnverifiedThe relationship between the standard error of Monte Carlo estimation and the number of simulations N is that the standard error is proportional to 1/√N.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1083976API
curl https://kongchang.com/api/v1/knowledge/claims/1083976MCP
get_claim(id=1083976)