Introduction
We have observed significant advancements in land cover classification during the past decades due to the availability of Earth Observation (EO) data and the rapid development of machine learning algorithms. There is a proliferation of advanced machine learning models that are highly predictive. Many remote sensing researchers and analysts use a model-centric approach that improves machine learning algorithms and overall accuracy. Although improving machine learning algorithms is good, limited and imbalanced training data impede machine learning models. In many cases, remote sensing researchers rarely communicate information on the quality of training data and their impact on model performance. Furthermore, most advanced machine learning models are opaque and difficult to interpret and explain.
Recently, it has become more critical than ever to understand and explain how machine learning models work since researchers use them to solve important land use and climate change challenges. As a result, experts have called the machine learning community to embrace a data-centric explainable machine learning approach. This approach focuses on improving the quality of training data and explaining how the models work. To improve the accuracy of machine learning models, remote sensing researchers and analysts should acquire high-quality training data, perform pre-processing, train the models and apply explainable machine learning techniques in an iterative process of model development.
Researchers have recently developed methods to address the complexity and explainability of machine learning models (Roscher et al. 2019, Apley and Zhu 2020). In addition, there have been calls to incorporate transparency and accountability in machine learning models. As a result, many researchers are working hard to develop explainable and interpretable machine learning models (Molnar 2019, Biecek and Burzykowski 2020). Explainable machine learning is difficult to define. In this course, explainable machine learning refers to the extent to which the underlying mechanism of a machine learning model can be explained (Biecek and Burzykowski 2020). That is, explainable machine learning models allow us (humans) to explain what the model learned and how it made predictions (post-hoc). Note this is different from interpretable machine learning (e.g., linear and logistic regression models), which refers to the extent to which a cause and effect are observed within a model (Molnar 2019).