Integration of process knowledge and statistical learning for the Dow data challenge problem

Qin, S. Joe; Guo, Siyi; Li, Zheyu; Chiang, Leo H.; Castillo, Iván; Braun, Birgit; Wang, Zhenyu

doi:10.1016/j.compchemeng.2021.107451

Cited by 17 publications

(11 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The one-to-one correspondence between the CG algorithm and the PLS1 algorithm is given in Algorithm 4. The equivalence in Step 2 can be derived using Equation (36). The equivalence between d i+1 and r i+1 in Step 5 can be shown as follows using Equation ( 27),…”

Section: Pls and The Conjugate Gradient Methodsmentioning

confidence: 99%

“…This result indicates that SDPLS explores more variations in X than PLS does while finding an optimal model to predict Y. online learning. 36 An important observation is that SDPLS achieves better R 2 values for both the training set and the testing set. Note that the PLS is a conjugate gradient method which finds the least squares optimum in fewer steps than SDPLS.…”

Section: Dow Chemical Impurity Data Modelingmentioning

confidence: 99%

“…There are outliers, missing values, and data from non‐standard operation models. In this article, we use the training dataset and testing dataset that were pre‐processed in Reference 36.…”

Section: Simulation and Industrial Case Studiesmentioning

confidence: 99%

“…The R 2 values of the testing set by SDPLS and PLS models are 0.484 and 0.475, respectively. The drop in the R 2 values for the testing set is due to the drifting effect in impurity, which can be compensated by online learning 36 . An important observation is that SDPLS achieves better R 2 values for both the training set and the testing set.…”

Section: Simulation and Industrial Case Studiesmentioning

confidence: 99%

“…The Dow data challenge problem was posted by Dow 35,36 to help academia test their algorithms with an industrial benchmark. the process flow diagram is shown in Figure 6, where the impurity of the primary SDPLS also explores the variance in X with its latent factors to predict Y.…”

Section: Dow Chemical Impurity Data Modelingmentioning

confidence: 99%

See 4 more Smart Citations

Partial least squares, steepest descent, and conjugate gradient for regularized predictive modeling

2022

View full text Add to dashboard Cite

In this article, we explore the connection of partial least squares (PLS) to other regularized regression algorithms including the Lasso and ridge regression, and consider a steepest descent alternative to the PLS algorithm. First, the PLS latent variable analysis is emphasized and formulated as a standalone procedure. The PLS connections to the conjugate gradient, Krylov space, and the Cayley-Hamilton theorem for matrix pseudo-inverse are explored based on known results in the literature. Comparison of PLS with the Lasso and ridge regression are given in terms of the different resolutions along the regularization paths, leading to an explanation of why PLS sometimes does not outperform the Lasso and ridge regression. As an attempt to increase resolutions along the regularization paths, a steepest descent PLS is formulated as a regularized regression alternative to PLS and is compared to other regularized algorithms via simulations and an industrial case study.

show abstract

Section: Pls and The Conjugate Gradient Methodsmentioning

confidence: 99%

Section: Dow Chemical Impurity Data Modelingmentioning

confidence: 99%

Section: Simulation and Industrial Case Studiesmentioning

confidence: 99%

Section: Simulation and Industrial Case Studiesmentioning

confidence: 99%

Section: Dow Chemical Impurity Data Modelingmentioning

confidence: 99%

See 3 more Smart Citations

Partial least squares, steepest descent, and conjugate gradient for regularized predictive modeling

2022

View full text Add to dashboard Cite

show abstract

A hybrid science‐guided machine learning approach for modeling chemical processes: A review

Sharma

Liu

2022

AIChE Journal

View full text Add to dashboard Cite

This study presents a broad perspective of hybrid process modeling combining the scientific knowledge and data analytics in bioprocessing and chemical engineering with a science-guided machine learning (SGML) approach. We divide the approach into two major categories: ML complements science, and science complements ML. We review the literature relating to the hybrid SGML approach, and propose a systematic classification of hybrid SGML models. For applying ML to improve science-based models, we present expositions of direct serial and parallel hybrid modeling and their combinations, inverse modeling, reduced-order modeling, quantifying uncertainty in the process and even discovering governing equations of the process model. For applying scientific principles to improve ML models, we discuss the science-guided design, learning and refinement. For each subcategory, we identify its requirements, strengths, and limitations, together with their published and potential applications. We also present several examples to illustrate different hybrid SGML methodologies for modeling chemical processes.

show abstract

Towards artificial intelligence at scale in the chemical industry

et al. 2022

Self Cite

View full text Add to dashboard Cite

In the Industry 4.0 era, the chemical industry is embracing broad adoption of artificial intelligence (AI) and machine learning (ML) methods. This article provides a holistic view of how the industry is transforming digitally towards AI at scale. First, a historical perspective on how the industry used AI to aid humans in better decision‐making is shown. Then state‐of‐the‐art AI research addressing industrial needs on reliability and safety, process optimization, supply chain, material discovery, and reaction engineering is highlighted. Finally, a vision of the plant of the future is illustrated with critical components of AI‐ready culture, model life cycle management, and renewed role of humans in chemical manufacturing.

show abstract

Integration of process knowledge and statistical learning for the Dow data challenge problem

Cited by 17 publications

References 28 publications

Partial least squares, steepest descent, and conjugate gradient for regularized predictive modeling

Partial least squares, steepest descent, and conjugate gradient for regularized predictive modeling

A hybrid science‐guided machine learning approach for modeling chemical processes: A review

Towards artificial intelligence at scale in the chemical industry

Contact Info

Product

Resources

About