☆ 4.7 Review

Explainable AI for earth observation: A review including societal and regulatory perspectives

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION (2022)

Related references

Note: Only part of the references are listed.

Article Computer Science, Artificial Intelligence

Embedding artificial intelligence in society: looking beyond the EU AI master plan using the culture cycle

Simone Borsci et al.

Summary: The article reviews the EU Commission's whitepaper on Artificial Intelligence and highlights potential conflicts with current societal, technical, and methodological constraints. The lack of a coherent EU vision and methods to support sustainable AI diffusion are identified as main obstacles. The article recommends complementary rules and compensatory mechanisms to avoid market fragmentation, as well as research to address technical and methodological open questions for the sustainable development of human-AI co-action.

AI & SOCIETY (2023)

Add to Collection

Article Geochemistry & Geophysics

Integrating Domain Knowledge in Data-Driven Earth Observation With Process Convolutions

Daniel Heestermans Svendsen et al.

Summary: The modeling of Earth observation data is a challenging problem that can be approached by either mechanistic or data-driven methods. This article proposes a hybrid learning scheme that combines the advantages of mechanistic models and machine learning methods, using Gaussian process convolution models. The proposed methodology efficiently addresses the limitations of traditional approaches in Earth observation modeling.

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING (2022)

Add to Collection

Article Agriculture, Multidisciplinary

An explainable XGBoost model improved by SMOTE-ENN technique for maize lodging detection based on multi-source unmanned aerial vehicle images

Liang Han et al.

Summary: This study uses machine learning models based on remote sensing images to detect crop lodging. The study uses Synthetic Minority Oversampling Technique and Edited Nearest Neighbors to handle imbalanced datasets, and proposes the SMOTE-ENN-XGBoost model for identifying maize lodging at the plot scale. SHapley Additive exPlanations approach is employed to interpret the features that determine lodging classification and activity prediction. The results suggest that canopy structure, spectral, and textural features should be considered simultaneously for accurate detection of crop lodging in crop breeding programs.

COMPUTERS AND ELECTRONICS IN AGRICULTURE (2022)

Add to Collection

Article Medicine, General & Internal

Applications of Explainable Artificial Intelligence in Diagnosis and Surgery

Yiming Zhang et al.

Summary: In recent years, artificial intelligence has shown promise in medicine, but lack of explainability limits its clinical applications. Explainable artificial intelligence (XAI) has been developed to overcome this limitation by providing both decision-making and explanations. This review surveys the recent trends in medical diagnosis and surgical applications using XAI and summarizes the methods, challenges, and future research directions.

DIAGNOSTICS (2022)

Add to Collection

Article Biodiversity Conservation

Determining the contribution of environmental factors in controlling dust pollution during cold and warm months of western Iran using different data mining algorithms and game theory

Zohre Ebrahimi-Khusfi et al.

Summary: The study used DNN models and game theory to predict dust pollution in arid regions of Iran, with findings showing that wind speed, soil heat flux, and precipitation are the most important controlling factors for dust storm occurrence. Results indicated that the DNN model was more accurate in predicting dust storm index compared to other traditional machine learning algorithms, potentially aiding in mitigating the impact of dust pollution on residents in western Iran.

ECOLOGICAL INDICATORS (2021)

Add to Collection

Article Geochemistry & Geophysics

Crop Yield Estimation and Interpretability With Gaussian Processes

Laura Martinez-Ferrer et al.

Summary: The study introduces the use of Gaussian processes for the estimation and understanding of crop development and yield using multisensor satellite observations and meteorological data, obtaining noticeable gains in accuracy.

IEEE GEOSCIENCE AND REMOTE SENSING LETTERS (2021)

Add to Collection

Article Parasitology

Modeling dengue vector population with earth observation data and a generalized linear model

Oladimeji Mudele et al.

Summary: Mosquitoes, particularly the Ae. aegypti species, play a critical role in transmitting diseases like Zika, Chikungunya, and Dengue. Understanding the environmental drivers of vector populations, such as temperature, moisture, precipitation, and vegetation, is crucial for effective public health interventions. By utilizing earth observation data and a Generalized Linear Model, researchers were able to model the relationship between mosquito populations and environmental conditions, providing an advancement in qualitative and explainable epidemiological risk modeling in urban environments.

ACTA TROPICA (2021)

Add to Collection

Article Agriculture, Multidisciplinary

Machine learning for large-scale crop yield forecasting

Dilli Paudel et al.

Summary: Many studies have applied machine learning to crop yield prediction with a focus on specific case studies, while large-scale operational systems like MCYFS do not use this method. By combining agronomic principles with machine learning, a baseline for large-scale crop yield forecasting was built in this research, emphasizing correctness, modularity, and reusability. The baseline showed comparable performance to MCYFS in early season predictions but has room for improvement with the addition of new data sources and predictive features.

AGRICULTURAL SYSTEMS (2021)

Add to Collection

Article Environmental Sciences

A Spatial-Temporal Interpretable Deep Learning Model for improving interpretability and predictive accuracy of satellite-based PM2.5

Xing Yan et al.

Summary: The development of a new Spatial-Temporal Interpretable Deep Learning Model (SIDLM) has been detailed in this study to enhance the interpretability and predictive accuracy of satellite-based PM2.5 measurements. The SIDLM demonstrated higher accuracy than five machine learning inversion methods and showed strong influence of certain districts on PM2.5 concentrations in urban areas such as Beijing. Overall, the new model has promising applications in deep learning-based predictions and spatiotemporal analysis of other earth observation data.

ENVIRONMENTAL POLLUTION (2021)

Add to Collection

Article Chemistry, Multidisciplinary

Integrating Image Quality Enhancement Methods and Deep Learning Techniques for Remote Sensing Scene Classification

Sheng-Chieh Hung et al.

Summary: By integrating technologies such as noise reduction, sharpening, and color adjustment, this research aims to enhance the quality of remote sensing images and benefit the extraction of image features from deep learning models, thereby improving classification efficiency.

APPLIED SCIENCES-BASEL (2021)

Add to Collection

Article Parasitology

Data-driven and interpretable machine-learning modeling to explore the fine-scale environmental determinants of malaria vectors biting rates in rural Burkina Faso

Paul Taconet et al.

Summary: This study was conducted in the rural health district of Diebougou, Burkina Faso, to explore the environmental determinants of human-biting activity in malaria vectors. The findings revealed significant correlations between meteorological and landscape variables and biting rates of the vectors, and identified complex associations through machine-learning models, providing new insights into the bio-ecology of malaria vectors in the area.

PARASITES & VECTORS (2021)

Add to Collection

Article Multidisciplinary Sciences

Multivariate random forest prediction of poverty and malnutrition prevalence

Chris Browne et al.

Summary: This research demonstrates how interpretable random forest models can produce estimates of correlated malnutrition and poverty prevalence measures using free, regularly updated, georeferenced data. The study found that the predictive accuracy of the model is broadly comparable to prior studies using proprietary data and deep or transfer learning methods, for both poverty mapping and early warning systems tasks.

PLOS ONE (2021)

Add to Collection

Article Meteorology & Atmospheric Sciences

Classifying precipitation from GEO satellite observations: Diagnostic model

Shruti A. Upadhyaya et al.

Summary: The study uses ML model to derive predictors from ABI sensor on GOES-16 satellite to study precipitation processes, finding satellite observations are crucial in distinguishing different precipitation types and different predictors are important for different types. It is recommended to combine heritage water vapor channel with infrared channel for improved precipitation classification accuracy.

QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY (2021)

Add to Collection

Article Chemistry, Analytical

Earthquake-Induced Building-Damage Mapping Using Explainable AI (XAI)

Sahar S. Matin et al.

Summary: Building-damage mapping using remote sensing images is crucial for providing information to first responders after earthquakes. The increasing interest in automatic post-earthquake building-damage maps through AI frameworks shows promise, but reliability issues exist that require further research and improvement.

SENSORS (2021)

Add to Collection

Article Chemistry, Analytical

Urban Vegetation Mapping from Aerial Imagery Using Explainable AI (XAI)

Abolfazl Abdollahi et al.

Summary: This study introduces a machine learning model based on deep neural networks for vegetation cover classification, incorporating explainable artificial intelligence techniques to help interpret the model's output. By testing with spectral and textural features, the model achieved an overall accuracy of 94.44% for vegetation cover mapping.

SENSORS (2021)

Add to Collection

Article Computer Science, Information Systems

A Fully Automatic, Interpretable and Adaptive Machine Learning Approach to Map Burned Area from Remote Sensing

Daniela Stroppiana et al.

Summary: The paper presents a fully automatic algorithm approach that combines knowledge-driven and data-driven elements to map burned areas from remote sensing. It utilizes active fire points to train a fusion function and applies a Region Growing algorithm to generate the final burned area map. The fusion function, an Ordered Weighted Averaging operator, is learned through machine learning and characterized by measures of pessimism/optimism and democracy/monarchy.

ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION (2021)

Add to Collection

Article Geochemistry & Geophysics

Toward a Collective Agenda on AI for Earth Science Data Analysis

Devis Tuia et al.

Summary: The fields of geosciences, remote sensing, and artificial intelligence have become closer in recent years, offering great opportunities to improve modeling and understanding of the Earth system. However, there is a need to challenge disciplinary comfort zones and inspire researchers to collaborate for real advancements in these areas.

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE (2021)

Add to Collection

Article Engineering, Electrical & Electronic

SAR Target Classification Based on Integration of ASC Parts Model and Deep Learning Algorithm

Sijia Feng et al.

Summary: This paper introduces two main methods for automatic target recognition of synthetic aperture radar (SAR) images - traditional machine learning and deep learning, and proposes a new target classification method that combines the advantages of both methods. Experimental results demonstrate the superiority of this method under complex conditions, and the analysis of target key parts with the part occlusion method contributes to the interpretability of the deep learning network.

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING (2021)

Add to Collection

Article Environmental Sciences

Bayesian U-Net: Estimating Uncertainty in Semantic Segmentation of Earth Observation Images

Clement Dechesne et al.

Summary: This work explores uncertainty estimations in semantic segmentation using a Bayesian deep learning method based on Monte Carlo Dropout, achieving high accuracy in semantic segmentation on several datasets with the U-Net architecture. The model also derives uncertainty maps, which allow for a more sound qualitative evaluation of the segmentation results and provide valuable information for improving reference databases.

REMOTE SENSING (2021)

Add to Collection

Article Plant Sciences

Explainable machine learning models of major crop traits from satellite-monitored continent-wide field trial data

Saul Justin Newman et al.

Summary: Despite the abundance of agricultural data, finding connections to crop yields remains challenging. Machine learning models can accurately predict crop behavior and reveal complex interactions, highlighting the powerful role of data in the future of food.

NATURE PLANTS (2021)

Add to Collection

Article Engineering, Electrical & Electronic

MIIDAPS-AI: An Explainable Machine-Learning Algorithm for Infrared and Microwave Remote Sensing and Data Assimilation Preprocessing-Application to LEO and GEO Sensors

Eric S. Maddy et al.

Summary: The MIIDAPS-AI algorithm utilizes advanced AI techniques to process satellite remote sensing data and generate vertical profiles and additional products such as temperature, moisture, surface, and cloud parameters. Compared to traditional algorithms, MIIDAPS-AI algorithm maintains accuracy while being more efficient, with faster computing speeds and cost savings.

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING (2021)

Add to Collection

Article Computer Science, Artificial Intelligence

Explainable AI for the Choquet Integral

Bryce J. Murray et al.

Summary: The modern era of machine learning emphasizes data-driven solutions, but explainability remains a challenge, as most machine learning solutions are black boxes. This paper focuses on data/information fusion, explores four XAI questions to enhance the accuracy of machine learning, and validates with experiments to achieve better future solutions.

IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE (2021)

Add to Collection

Article Remote Sensing

Earth observation tools and services to increase the effectiveness of humanitarian assistance

Stefan Lang et al.

EUROPEAN JOURNAL OF REMOTE SENSING (2020)

Add to Collection

Article Environmental Sciences

Estimating and understanding crop yields with explainable deep learning in the Indian Wheat Belt

Aleksandra Wolanin et al.

ENVIRONMENTAL RESEARCH LETTERS (2020)

Add to Collection

Article Computer Science, Information Systems

Measuring Impacts of Urban Environmental Elements on Housing Prices Based on Multisource Data-A Case Study of Shanghai, China

Liujia Chen et al.

ISPRS INTERNATIONAL JOURNAL OF GEO-INFORMATION (2020)

Add to Collection

Article Environmental Sciences

Canopy Height Estimation Using Sentinel Series Images through Machine Learning Models in a Mangrove Forest

Sujit Madhab Ghosh et al.

REMOTE SENSING (2020)

Add to Collection

Article Environmental Sciences

Crop Type Classification Using Fusion of Sentinel-1 and Sentinel-2 Data: Assessing the Impact of Feature Selection, Optical Data Availability, and Parcel Sizes on the Accuracies

Aiym Orynbaikyzy et al.

REMOTE SENSING (2020)

Add to Collection

Article Environmental Sciences

Estimating spatial and temporal variation in ocean surface pCO2 in the Gulf of Mexico using remote sensing and machine learning techniques

Zhiyi Fu et al.

SCIENCE OF THE TOTAL ENVIRONMENT (2020)

Add to Collection

Article Computer Science, Artificial Intelligence

Enabling Explainable Fusion in Deep Learning With Fuzzy Integral Neural Networks

Muhammad Aminul Islam et al.

IEEE TRANSACTIONS ON FUZZY SYSTEMS (2020)

Add to Collection

Article Computer Science, Information Systems

Explainable Machine Learning for Scientific Insights and Discoveries

Ribana Roscher et al.

IEEE ACCESS (2020)

Add to Collection

Article Computer Science, Information Systems

Modeling the Temporal Population Distribution of Ae. aegypti Mosquito Using Big Earth Observation Data

Oladimeji Mudele et al.

IEEE ACCESS (2020)

Add to Collection

Article Computer Science, Artificial Intelligence

Explanation in artificial intelligence: Insights from the social sciences

Tim Miller

ARTIFICIAL INTELLIGENCE (2019)

Add to Collection

Article Multidisciplinary Sciences

Unmasking Clever Hans predictors and assessing what machines really learn

Sebastian Lapuschkin et al.

NATURE COMMUNICATIONS (2019)

Add to Collection

Review Geography, Physical

Deep learning in remote sensing applications: A meta-analysis and review

Lei Ma et al.

ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING (2019)

Add to Collection

Article Multidisciplinary Sciences

Definitions, methods, and applications in interpretable machine learning

W. James Murdoch et al.

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA (2019)

Add to Collection

Article Computer Science, Information Systems

Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)

Amina Adadi et al.

IEEE ACCESS (2018)

Add to Collection

Article Environmental Sciences

One-Dimensional Convolutional Neural Network Land-Cover Classification of Multi-Seasonal Hyperspectral Imagery in the San Francisco Bay Area, California

Daniel Guidici et al.

REMOTE SENSING (2017)

Add to Collection

Article Geochemistry & Geophysics

Deep Learning in Remote Sensing

Xiao Xiang Zhu et al.

IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE (2017)

Add to Collection

Article Computer Science, Artificial Intelligence

European Union Regulations on Algorithmic Decision Making and a Right to Explanation

Bryce Goodman et al.

AI MAGAZINE (2017)

Add to Collection

Review Geography, Physical

Random forest in remote sensing: A review of applications and future directions

Mariana Belgiu et al.

ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING (2016)

Add to Collection

Article Geochemistry & Geophysics

A machine learning approach to geochemical mapping

Charlie Kirkwood et al.

JOURNAL OF GEOCHEMICAL EXPLORATION (2016)

Add to Collection

Article Social Sciences, Interdisciplinary

How the machine 'thinks': Understanding opacity in machine learning algorithms

Jenna Burrell

BIG DATA & SOCIETY (2016)

Add to Collection

Article Remote Sensing

Multi-scale object-based image analysis and feature selection of multi-sensor earth observation imagery using random forests

Dennis C. Duro et al.

INTERNATIONAL JOURNAL OF REMOTE SENSING (2012)

Add to Collection

Article Geography, Physical

Regional mapping of soil parent material by machine learning based on point data

Marine Lacoste et al.

GEOMORPHOLOGY (2011)

Add to Collection

Article Operations Research & Management Science

A disaster-severity assessment DSS comparative analysis

J. Tinguaro Rodriguez et al.

OR SPECTRUM (2011)

Add to Collection

Article Computer Science, Artificial Intelligence

Gaining insight through case-based explanation

Conor Nugent et al.

JOURNAL OF INTELLIGENT INFORMATION SYSTEMS (2009)

Add to Collection

Article Remote Sensing

A technique for feature selection in multiclass problems

L Bruzzone et al.

INTERNATIONAL JOURNAL OF REMOTE SENSING (2000)

Add to Collection

© Peeref 2019-2024. All rights reserved.