ContinuAR: Continuous Autoregression For Infinite-Fidelity Fusion

Authors: WEI XING, Yuxin Wang, Zheng Xing

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

Reproducibility Variable Result LLM Response
Research Type Experimental To assess Continu AR, we compare it with (1) AR [3], (2) Res GP [12], (3) MF-BNN2 [33], and (4) IFC3 [8], which are the most closely related SOTA methods for high-dimensional multi-fidelity fusion, particularly with infinite fidelities. ... All experiments are run on a workstation with an AMD 5950x CPU, Nvidia RTX3090 GPU, and 32 GB RAM. All experiments are repeated five times with different random seeds, and the mean performance and its standard deviation are reported.
Researcher Affiliation Collaboration Wei W. Xing School of Mathematics and Statistics, University of Sheffield Hicks Building, Hounsfield Rd, Sheffield, UK, S3 7RH w.xing@sheffield.ac.uk Yuxin Wang Department of Statistics and Data Science National Univerisity of Singapore 21 Lower Kent Ridge Road, Singapore, 119077. yuxinwang@u.nus.edu Zheng Xing Graphics& Computing Department Rockchip Electronics Co., Ltd Fuzhou, China, 350003. zheng.xing@rock-chips.com
Pseudocode No The paper does not contain any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper mentions links for MF-BNN2 (https://github.com/shib0li/DNN-MFBO) and IFC3 (https://github.com/shib0li/Infinite-Fidelity-Coregionalization) but does not provide concrete access to its own source code for the Continu AR method.
Open Datasets No The paper describes how the data was generated through simulations (e.g., Heat, Burgers, and Poisson's equations, Top OP, PNA) but does not provide concrete access information (link, DOI, repository, or citation) for these specific datasets.
Dataset Splits No The paper describes training data selection and testing, but does not explicitly provide details for a separate validation set or specific train/validation/test splits.
Hardware Specification Yes All experiments are run on a workstation with an AMD 5950x CPU, Nvidia RTX3090 GPU, and 32 GB RAM.
Software Dependencies No The paper states: 'Continu AR, AR, and Res GP are implemented using Pytorch.' It does not specify version numbers for Pytorch or any other software dependencies.
Experiment Setup Yes All GP-based models are trained with 200 iterations whereas MF-BNN and IFC with 1000 iterations to reach convergence.