Minimax-Optimal Location Estimation

Authors: Shivam Gupta, Jasper Lee, Eric Price, Paul Valiant

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We compare various location estimation methods on synthetic data from a fairly simple, but irregular, piecewise linear distribution (Figure 3(a)). We set n = 10 and aim for 90% confidence intervals. In Figure 3(b), we plot the CDF of the point error produced by the MLE, the 0.3-smoothed MLE, and our two algorithms (Algorithm 1 and Algorithm 4).
Researcher Affiliation Academia Shivam Gupta The University of Texas at Austin shivamgupta@utexas.edu Jasper C.H. Lee University of Wisconsin Madison jasper.lee@wisc.edu Eric Price The University of Texas at Austin ecprice@cs.utexas.edu Paul Valiant Purdue University pvaliant@gmail.com
Pseudocode Yes Algorithm 1 The algorithm At for a fixed estimation accuracy ϵ [...] Algorithm 2. Consider the optimal failure probability δ as a function of the estimation accuracy ϵ. [...] Algorithm 3 Estimator minimizing RA cos ρ + δA sin ρ for a given angle ρ [0, π/2] [...] Algorithm 4 (Minimax-optimal confidence-interval estimator). [...] Algorithm 5 Approximately computing ϵ from δ using binary search [...] Algorithm 6 Binary search for the optimal estimator through the slope angle ρ
Open Source Code No The paper does not provide any explicit statements about open-source code availability or links to code repositories.
Open Datasets No We compare various location estimation methods on synthetic data from a fairly simple, but irregular, piecewise linear distribution (Figure 3(a)).
Dataset Splits No The paper does not explicitly provide training/validation/test dataset splits or mention cross-validation. It uses 'synthetic data' for evaluation.
Hardware Specification No The paper does not provide any specific hardware details used for running the experiments.
Software Dependencies No The paper does not specify any software dependencies with version numbers.
Experiment Setup Yes We set n = 10 and aim for 90% confidence intervals.