https://github.com/KislayaRavi/MuDaFuGP

https://github.com/KislayaRavi/MuDaFuGP

Last synced: 11 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/KislayaRavi/MuDaFuGP
• Owner: KislayaRavi
• Created: 2022-10-26T15:46:55.000Z (about 2 years ago)
• Default Branch: main
• Last Pushed: 2024-07-23T13:21:55.000Z (4 months ago)
• Last Synced: 2024-08-01T16:45:46.756Z (3 months ago)
• Language: Jupyter Notebook
• Size: 2.55 MB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
![Logo](./docs/materials/logo/logo.png)

# MuDaFu GP

This project provides parameterized implementations of various multi-fidelity Gaussian Process Regression algorithms:

- Nonlinear autoregressive multi-fidelity Gaussian Processes [[1]](#1)

- Gaussian Processes with Data Fusion and Delays [[2]](#1)

Underfitted models can be efficiently improved using an entropy reduction method called Adaptation.

This repo also provides a Polynomial Chaos Expansion implementation, which can be performed on the mean prediction functions of MFGPs.

Linking PCE and multi-fidelity models leads to equal precisions as direct PCE but needs much less high-fidelity model evaluations.

This saves a significant amount of computation effort.

This library also supports applications:

- Multifidelity forward uncertainity quantification

- Multifidelity No-Turn Sampling

- Multifildelity Bayesian Optimization 

# References

[1] 

Perdikaris, Paris, et al. "Nonlinear information fusion algorithms for data-efficient multi-fidelity modelling." Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 473.2198 (2017): 20160751.

[2] 

S. Lee, F. Dietrich, G. E. Karniadakis, and I. G. Kevrekidis. “Linking Gaussianprocess regression with data-driven manifold embeddings for nonlinear datafusion.” In:Interface focus9.3 (2019), p. 20180083.