Sequential Kernel Goodness-of-fit Testing

Authors: Zhengyu Zhou, Weiwei Liu

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct experiments to demonstrate the adaptability of our sequential test across varying difficulty levels of problems while maintaining control over type-I errors. This section describes the experiments that demonstrate our tests capacity to adapt to a problem s unknown difficulty while maintaining type-I error control.
Researcher Affiliation Academia 1School of Computer Science, National Engineering Research Center for Multimedia Software, Institute of Artificial Intelligence and Hubei Key Laboratory of Multimedia and Network Communication Engineering, Wuhan University, Wuhan, China. Correspondence to: Weiwei Liu <liuweiwei863@gmail.com>.
Pseudocode Yes Algorithm 1 Online Newton Step (ONS) strategy for selecting betting fractions, Algorithm 2 KSD-based SKGT, Algorithm 3 KDSD-based SKGT, Algorithm 4 bd-KSD-based SKGT
Open Source Code No The paper does not provide any explicit statements about releasing source code or links to a code repository.
Open Datasets No The paper describes how synthetic data was generated for simulations (e.g., "Zt iid q", "samples from a Student s t distribution", "samples from an Ising model"), but it does not provide access information or citations for any public dataset.
Dataset Splits No The paper conducts statistical goodness-of-fit tests and does not involve training machine learning models with explicit training, validation, and testing data splits.
Hardware Specification No The paper does not provide any specific details regarding the hardware used for running the experiments.
Software Dependencies No The paper mentions statistical methods and algorithms used (e.g., Gaussian kernel, Metropolis algorithm, wild bootstrap), but it does not list any specific software dependencies with version numbers (e.g., programming languages, libraries, or specialized packages).
Experiment Setup Yes We consider 11 β: β {0.5, 0.55, . . . , 1.0} values, and for each β we repeat the simulation 100 times. We use Gaussian kernel k(x, y) = exp( |x y|2 /2) for all testing procedures. We consider a periodic 10-by-10 lattice, with d = 100 random variables. We focus on the ferromagnetic setting and set θij = 1/T, where T is the temperature of the system. For T0 = 5 and various values of T, we test the hypotheses H0 : T = T0 vs. H1 : T = T0 using data samples drawn from the model under temperature T. We consider different values of v {0, 0.3, 0.6, 0.9}.