Minimax Lower Bounds for Estimating Distributions on Low-dimensional Spaces

Authors: Saptarshi Chakraborty

TMLR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical This paper demonstrates that the minimax rate for estimating unknown distributions in the β-Hölder IPM on M scales as Ω n β d M δ n 1/2, where d M is the lower Minkowski dimension of M. Thus if the low-dimensional structure M is regular in the Minkowski sense, i.e. d M = d M, GANs are roughly minimax optimal in estimating distributions on M. Further, the paper shows that the minimax estimation rate in the p-Wasserstein metric scales as Ω n 1 d M δ n 1/(2p) .
Researcher Affiliation Academia Saptarshi Chakraborty EMAIL Department of Statistics University of California, Berkeley
Pseudocode No The paper focuses on theoretical analysis and mathematical proofs, such as "3 Proof of the Main Result (Theorem 7)", without including any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not contain any statements about releasing source code, nor does it provide links to code repositories or supplementary materials with code.
Open Datasets No The paper is a theoretical work focusing on minimax lower bounds for estimating distributions, and thus does not utilize or provide access information for any specific open datasets for experimental validation.
Dataset Splits No The paper does not describe any experimental setup involving datasets or their splits, as it focuses on theoretical analysis.
Hardware Specification No The paper is a theoretical work and does not describe any experiments that would require specific hardware specifications.
Software Dependencies No The paper is a theoretical work and does not mention any specific software dependencies or versions required for replication.
Experiment Setup No The paper is a theoretical analysis of minimax lower bounds and does not provide details on experimental setup, hyperparameters, or training configurations.