Knowledge-Guided Machine Learning

Karpatne, Anuj; Kannan, Ramakrishnan; Kumar, Vipin

doi:10.1201/9781003143376

Cited by 31 publications

(17 citation statements)

References 132 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To highlight how the hybrid MCL model performs in capturing key physical limnological lake characteristics, we ran the hybrid MCL model against the process‐based model, the deep learning with no process, and the deep learning with no modularization in a time period that was not used previously for neither training nor testing. To avoid confusion, we are using the following nomenclature in this study: KGML : Knowledge‐Guided Machine Learning, a modeling paradigm that aims to combine process‐based knowledge and modeling with deep learning models (Karpatne et al., 2017, 2022). MCL : Modular Compositional Learning, a KGML methodology in which the overall model is decomposed into modular sub‐aspects, each modular sub‐aspect can be a deep learning model or a process‐based model (Karpatne et al., 2017).…”

Section: Methodsmentioning

confidence: 99%

Modular Compositional Learning Improves 1D Hydrodynamic Lake Model Performance by Merging Process‐Based Modeling With Deep Learning

Ladwig,

Daw,

Albright

et al. 2023

J Adv Model Earth Syst

View full text Add to dashboard Cite

Hybrid Knowledge‐Guided Machine Learning (KGML) models, which are deep learning models that utilize scientific theory and process‐based model simulations, have shown improved performance over their process‐based counterparts for the simulation of water temperature and hydrodynamics. We highlight the modular compositional learning (MCL) methodology as a novel design choice for the development of hybrid KGML models in which the model is decomposed into modular sub‐components that can be process‐based models and/or deep learning models. We develop a hybrid MCL model that integrates a deep learning model into a modularized, process‐based model. To achieve this, we first train individual deep learning models with the output of the process‐based models. In a second step, we fine‐tune one deep learning model with observed field data. In this study, we replaced process‐based calculations of vertical diffusive transport with deep learning. Finally, this fine‐tuned deep learning model is integrated into the process‐based model, creating the hybrid MCL model with improved overall projections for water temperature dynamics compared to the original process‐based model. We further compare the performance of the hybrid MCL model with the process‐based model and two alternative deep learning models and highlight how the hybrid MCL model has the best performance for projecting water temperature, Schmidt stability, buoyancy frequency, and depths of different isotherms. Modular compositional learning can be applied to existing modularized, process‐based model structures to make the projections more robust and improve model performance by letting deep learning estimate uncertain process calculations.

show abstract

Section: Methodsmentioning

confidence: 99%

Modular Compositional Learning Improves 1D Hydrodynamic Lake Model Performance by Merging Process‐Based Modeling With Deep Learning

Ladwig,

Daw,

Albright

et al. 2023

J Adv Model Earth Syst

View full text Add to dashboard Cite

show abstract

“…Fourth, spatiotemporal problems are often highly interdisciplinary and require the integration of deep learning with physical knowledge [29]. This is because machine learning problems are often under-constrained.…”

Section: Unique Challengesmentioning

confidence: 99%

Analysis of Computer Information Processing Technology Under the Background of Intelligent Big Data

Jiang¹,

Zhi-ling²

2020

2020 International Conference on Computer Vision, Image and Deep Learning (CVIDL)

View full text Add to dashboard Cite

With advancements in GPS, remote sensing, and computational simulation, an enormous volume of spatiotemporal data is being collected at an increasing speed from various application domains, spanning Earth sciences, agriculture, smart cities, and public safety. Such emerging geospatial and spatiotemporal big data, coupled with recent advances in deep learning technologies, foster new opportunities to solve problems that have not been possible before. For instance, remote sensing researchers can potentially train a foundation model using Earth imagery big data for numerous land cover and land use modeling tasks. Coastal modelers can train AI surrogates to speed up numerical simulations. However, the distinctive characteristics of spatiotemporal big data pose new challenges for deep learning technologies. This vision paper introduces various types of spatiotemporal big data, discusses new research opportunities in the realm of deep learning applied to spatiotemporal big data, lists the unique challenges, and identifies several future research needs.

show abstract

“…The approaches of Wekesa et al and Wang et al can be classified as hybrid ML approaches, which aim to capture "the best of both worlds" by combining the predictive powers of data-driven approaches of ML with the interpretability of theory-based models such as mechanistic or knowledge-based models. These types of models have been applied to several domains, most notably in physics-informed ML models 31 , and with respect to computational biology have been applied through several frameworks. ML models have been integrated with mechanistic models of metabolism to determine kinetic parameters and predict downstream metabolic effects 32 .…”

Section: Introductionmentioning

confidence: 99%

A hybrid machine learning framework for functional annotation applied to mitochondrial glutathione metabolism and transport in cancers.

Kennedy,

Sandhu,

Harper

et al. 2023

Preprint

View full text Add to dashboard Cite

Alterations of metabolism are a well appreciated hallmark of many cancers, including changes in mitochondrial and glutathione (GSH) metabolism. Mitochondrial GSH (mGSH) transport is a poorly characterized aspect of GSH metabolism, which we investigate in the context of cancer. Here, we apply hybrid machine learning (ML) models and leverage cancer cell line multi-omics data and other biological knowledge data as features, to uncover potential candidates involved in mGSH metabolism and membrane transport in cancers. These models predict the known mGSH transporter SLC25A39 but not SLC25A40 as being highly probably related to GSH metabolism in cancers. Additionally, SLC25A24 and SLC25A37 along with the poorly characterized members of the SLC25 family, SLC25A43 and SLC25A50, are predicted to be associated. Similarities in potential substrate binding regions are found between SLC25A39 and the candidates SLC25A24 and A43. These findings have implications for a better understanding of cancer cell metabolism and novel therapeutic targets.

show abstract

Knowledge-Guided Machine Learning

Cited by 31 publications

References 132 publications

Modular Compositional Learning Improves 1D Hydrodynamic Lake Model Performance by Merging Process‐Based Modeling With Deep Learning

Modular Compositional Learning Improves 1D Hydrodynamic Lake Model Performance by Merging Process‐Based Modeling With Deep Learning

Analysis of Computer Information Processing Technology Under the Background of Intelligent Big Data

A hybrid machine learning framework for functional annotation applied to mitochondrial glutathione metabolism and transport in cancers.

Contact Info

Product

Resources

About