On the Estimation of Gaussian Mixture Copula Models

Authors: Ashutosh Tewari

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

Reproducibility Variable Result LLM Response
Research Type Experimental The main claims of this work are supported by empirical evidence gathered on synthetic and realworld datasets. and Section 6. Experimental Results This section provides empirical evidence to support the claims made in this paper using synthetic and real-world datasets.
Researcher Affiliation Industry 1Amazon Mechatronics, Seattle, USA. Correspondence to: Ashutosh Tewari <ashutosh80@gmail.com>.
Pseudocode No The paper describes algorithmic procedures (e.g., EM algorithm, numerical schemes for quantile computation) but does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks with structured steps.
Open Source Code Yes The contributed GMCM code (Tewari, 2023) employs the parallel implementation of Chandrupatla s algorithm via the Tensor Flow Probability package. and Code used in this analysis is available here (Tewari, 2023).
Open Datasets Yes Finally, GMCM is learned on several density estimation benchmarks from the UCI repository, after following the pre-processing step described in (Papamakarios et al., 2017).
Dataset Splits Yes This number was ascertained by a grid search over {10,20,30,40,50} and tracking the likelihood of validation datasets.
Hardware Specification No The paper mentions 'CPU time' in relation to computational complexity but does not specify any particular CPU model, GPU, or other hardware specifications used for the experiments.
Software Dependencies No The paper mentions the use of 'Python' and the 'Tensor Flow-Probability package' but does not specify their version numbers or other key software dependencies with specific versions.
Experiment Setup Yes The maximization is carried out using the Adam optimizer (Kingma & Ba, 2017) with the default learning rate of 1E 3. and the mixture models (GMCM and GMM) were instantiated with the 40 mixing components.