With the availability of data and computational technologies in the modern world, machine learning (ML) has emerged as a preferred methodology for data analysis and prediction. While ML holds great promise, the results from such models are not fully unreliable due to the challenges introduced by uncertainty. An ML model generates an optimal solution based on its training data. However, if the uncertainty in the data and the model parameters are not considered, such optimal solutions have a high risk of failure in actual world deployment. This paper surveys the different approaches used in ML to quantify uncertainty. The paper also exhibits the implications of quantifying uncertainty when using ML by performing two case studies with space physics in focus. The first case study consists of the classification of auroral images in predefined labels. In the second case study, the horizontal component of the perturbed magnetic field measured at the Earth’s surface was predicted for the study of Geomagnetically Induced Currents (GICs) by training the model using time series data. In both cases, a Bayesian Neural Network (BNN) was trained to generate predictions, along with epistemic and aleatoric uncertainties. Finally, the pros and cons of both Gaussian Process Regression (GPR) models and Bayesian Deep Learning (DL) are weighed. The paper also provides recommendations for the models that need exploration, focusing on space weather prediction.

No abstract

Most soil quality measurements have been limited to laboratory-based methods that suffer from time delay, high cost, intensive labor requirement, discrete data collection, and tedious sample pretreatment. Real-time continuous soil monitoring (RTCSM) possesses a great potential to revolutionize field measurements by providing first-hand information for continuously tracking variations of heterogeneous soil parameters and diverse pollutants in a timely manner and thus enable constant updates essential for system control and decision-making. Through a systematic literature search and comprehensive analysis of state-of-the-art RTCSM technologies, extensive discussion of their vital hurdles, and sharing of our future perspectives, this critical review bridges the knowledge gap of spatiotemporal uninterrupted soil monitoring and soil management execution. First, the barriers for reliable RTCSM data acquisition are elucidated by examining typical soil monitoring techniques (e.g., electrochemical and spectroscopic sensors). Next, the prevailing challenges of the RTCSM sensor network, data transmission, data processing, and personalized data management are comprehensively discussed. Furthermore, this review explores RTCSM data application for updating diverse strategies including high-fidelity soil process models, control methodologies, digital soil mapping, soil degradation, food security, and climate change mitigation. Finally, the significance of RTCSM implementation in agricultural and environmental fields is underscored through illuminating future directions and perspectives in this systematic review.

No abstract

No abstract