Soil Science-Informed Machine Learning
Machine learning (ML) applications in soil science have significantly increased over the past two decades, reflecting a growing trend towards data-driven research addressing soil security. This extensive application has mainly focused on enhancing predictions of soil properties, particularly soil or...
Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-12-01
|
| Series: | Geoderma |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S0016706124003239 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1846139905199046656 |
|---|---|
| author | Budiman Minasny Toshiyuki Bandai Teamrat A. Ghezzehei Yin-Chung Huang Yuxin Ma Alex B. McBratney Wartini Ng Sarem Norouzi Jose Padarian Rudiyanto Amin Sharififar Quentin Styc Marliana Widyastuti |
| author_facet | Budiman Minasny Toshiyuki Bandai Teamrat A. Ghezzehei Yin-Chung Huang Yuxin Ma Alex B. McBratney Wartini Ng Sarem Norouzi Jose Padarian Rudiyanto Amin Sharififar Quentin Styc Marliana Widyastuti |
| author_sort | Budiman Minasny |
| collection | DOAJ |
| description | Machine learning (ML) applications in soil science have significantly increased over the past two decades, reflecting a growing trend towards data-driven research addressing soil security. This extensive application has mainly focused on enhancing predictions of soil properties, particularly soil organic carbon, and improving the accuracy of digital soil mapping (DSM). Despite these advancements, the application of ML in soil science faces challenges related to data scarcity and the interpretability of ML models. There is a need for a shift towards Soil Science-Informed ML (SoilML) models that use the power of ML but also incorporate soil science knowledge in the training process to make predictions more reliable and generalisable. This paper proposes methodologies for embedding ML models with soil science knowledge to overcome current limitations. Incorporating soil science knowledge into ML models involves using observational priors to enhance training datasets, designing model structures which reflect soil science principles, and supervising model training with soil science-informed loss functions. The informed loss functions include observational constraints, coherency rules such as regularisation to avoid overfitting, and prior or soil-knowledge constraints that incorporate existing information about the parameters or outputs. By way of illustration, we present examples from four fields: digital soil mapping, soil spectroscopy, pedotransfer functions, and dynamic soil property models. We discuss the potential to integrate process-based models for improved prediction, the use of physics-informed neural networks, limitations, and the issue of overparametrisation. These approaches improve the relevance of ML predictions in soil science and enhance the models’ ability to generalise across different scenarios while maintaining soil science principles, transparency and reliability. |
| format | Article |
| id | doaj-art-b96fd0fe95a043a689a5f2030a96a07f |
| institution | Kabale University |
| issn | 1872-6259 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Geoderma |
| spelling | doaj-art-b96fd0fe95a043a689a5f2030a96a07f2024-12-06T05:12:42ZengElsevierGeoderma1872-62592024-12-01452117094Soil Science-Informed Machine LearningBudiman Minasny0Toshiyuki Bandai1Teamrat A. Ghezzehei2Yin-Chung Huang3Yuxin Ma4Alex B. McBratney5Wartini Ng6Sarem Norouzi7Jose Padarian8 Rudiyanto9Amin Sharififar10Quentin Styc11Marliana Widyastuti12School of Life and Environmental Sciences & Sydney Institute of Agriculture, The University of Sydney, NSW 2006, AustraliaEarth and Environmental Sciences Area, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USALife & Environmental Sciences Department, University of California, Merced, CA 95343, USASchool of Life and Environmental Sciences & Sydney Institute of Agriculture, The University of Sydney, NSW 2006, AustraliaNew South Wales Department of Climate Change, Energy, the Environment and Water, Parramatta, NSW 2150, AustraliaSchool of Life and Environmental Sciences & Sydney Institute of Agriculture, The University of Sydney, NSW 2006, AustraliaSchool of Life and Environmental Sciences & Sydney Institute of Agriculture, The University of Sydney, NSW 2006, AustraliaDepartment of Agroecology, Aarhus University, 8830 Tjele, DenmarkSchool of Life and Environmental Sciences & Sydney Institute of Agriculture, The University of Sydney, NSW 2006, AustraliaFaculty of Fisheries and Food Science, Universiti Malaysia Terengganu, 21030 Kuala Nerus, Terengganu, MalaysiaSchool of Life and Environmental Sciences & Sydney Institute of Agriculture, The University of Sydney, NSW 2006, AustraliaSchool of Life and Environmental Sciences & Sydney Institute of Agriculture, The University of Sydney, NSW 2006, AustraliaSchool of Life and Environmental Sciences & Sydney Institute of Agriculture, The University of Sydney, NSW 2006, AustraliaMachine learning (ML) applications in soil science have significantly increased over the past two decades, reflecting a growing trend towards data-driven research addressing soil security. This extensive application has mainly focused on enhancing predictions of soil properties, particularly soil organic carbon, and improving the accuracy of digital soil mapping (DSM). Despite these advancements, the application of ML in soil science faces challenges related to data scarcity and the interpretability of ML models. There is a need for a shift towards Soil Science-Informed ML (SoilML) models that use the power of ML but also incorporate soil science knowledge in the training process to make predictions more reliable and generalisable. This paper proposes methodologies for embedding ML models with soil science knowledge to overcome current limitations. Incorporating soil science knowledge into ML models involves using observational priors to enhance training datasets, designing model structures which reflect soil science principles, and supervising model training with soil science-informed loss functions. The informed loss functions include observational constraints, coherency rules such as regularisation to avoid overfitting, and prior or soil-knowledge constraints that incorporate existing information about the parameters or outputs. By way of illustration, we present examples from four fields: digital soil mapping, soil spectroscopy, pedotransfer functions, and dynamic soil property models. We discuss the potential to integrate process-based models for improved prediction, the use of physics-informed neural networks, limitations, and the issue of overparametrisation. These approaches improve the relevance of ML predictions in soil science and enhance the models’ ability to generalise across different scenarios while maintaining soil science principles, transparency and reliability.http://www.sciencedirect.com/science/article/pii/S0016706124003239Artificial IntelligenceProcess-based modelsPhysics Informed Neural NetworksInformed Machine LearningMechanistic modelsPedology |
| spellingShingle | Budiman Minasny Toshiyuki Bandai Teamrat A. Ghezzehei Yin-Chung Huang Yuxin Ma Alex B. McBratney Wartini Ng Sarem Norouzi Jose Padarian Rudiyanto Amin Sharififar Quentin Styc Marliana Widyastuti Soil Science-Informed Machine Learning Geoderma Artificial Intelligence Process-based models Physics Informed Neural Networks Informed Machine Learning Mechanistic models Pedology |
| title | Soil Science-Informed Machine Learning |
| title_full | Soil Science-Informed Machine Learning |
| title_fullStr | Soil Science-Informed Machine Learning |
| title_full_unstemmed | Soil Science-Informed Machine Learning |
| title_short | Soil Science-Informed Machine Learning |
| title_sort | soil science informed machine learning |
| topic | Artificial Intelligence Process-based models Physics Informed Neural Networks Informed Machine Learning Mechanistic models Pedology |
| url | http://www.sciencedirect.com/science/article/pii/S0016706124003239 |
| work_keys_str_mv | AT budimanminasny soilscienceinformedmachinelearning AT toshiyukibandai soilscienceinformedmachinelearning AT teamrataghezzehei soilscienceinformedmachinelearning AT yinchunghuang soilscienceinformedmachinelearning AT yuxinma soilscienceinformedmachinelearning AT alexbmcbratney soilscienceinformedmachinelearning AT wartining soilscienceinformedmachinelearning AT saremnorouzi soilscienceinformedmachinelearning AT josepadarian soilscienceinformedmachinelearning AT rudiyanto soilscienceinformedmachinelearning AT aminsharififar soilscienceinformedmachinelearning AT quentinstyc soilscienceinformedmachinelearning AT marlianawidyastuti soilscienceinformedmachinelearning |