Rose Nakibuule
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Rose Nakibuule has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Ggaliwango Marvin
Department of Computer Science, Makerere University, Kampala, Uganda
5 shared publications - Mercy Amiyo
Makerere University
2 shared publications - Amos Ojok
Makerere University
1 shared publication Christine Kikome
Makerere University
1 shared publication- Edward Kaboggoza
Makerere University, Uganda
1 shared publication Emmanuel Wagisha
Makerere University
1 shared publication
Latest Publications
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 157539-157567
Large Acoustic Models For Visual Scene Narration With Large Visual Embeddings
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 157539-157567
Open AccessLarge Acoustic Models For Visual Scene Narration With Large Visual Embeddings
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 157845-157866
Interpretable Attention-Based Mechanisms for Medical Waste Sorting With Meta Learning
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 157845-157866
Open AccessInterpretable Attention-Based Mechanisms for Medical Waste Sorting With Meta Learning
- 13 July 2025
- Published by Springer Nature
- p. 605-621
Interpretable Class Activation Maps for Vision-Based Deforestation Detection
- 13 July 2025
- Published by Springer Nature
- p. 605-621
Interpretable Class Activation Maps for Vision-Based Deforestation Detection
- Open Access
- 8 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 271-281
Transparent Multi-modelling for Climatic Geo-intelligence in Predicting Food Security
Millions of people in Uganda suffer from food insecurity as a result of the country’s scarce resources, extreme poverty, and high rate of childhood malnutrition. Women, who play a critical role in home management, are particularly vulnerable to food insecurity brought on by climate change. The vision is to foster a thriving agricultural sector in Uganda grounded in sustainability, with a mission to empower farmers, including women, through innovative tools and knowledge to adapt and prosper in a changing climate. This study employs cutting-edge technologies such as geospatial data analysis, machine learning, and environmental science to analyze weather patterns, soil conditions, and crop data. By providing personalized insights and recommendations to farmers, the research enables informed decision-making, optimizes resource management, and mitigates climate-related risks in agriculture. The research utilizes a multimodal architecture incorporating 13 traditional base machine learning models including Linear Regression, Lasso, Ridge, Elastic Net, Random Forest, K-Nearest Neighbors, Decision Tree, Support Vector Machine, Gaussian Process Regression, Gradient Boosting Regressor, XGBoost, LightGBM, and MLPRegressor. These base models are integrated into an ensemble approach using Stacking and Voting Regressors to harness collective intelligence and enhance predictive accuracy in optimizing agricultural productivity and resilience. The findings reveal that among the base models, LightGBM demonstrates superior performance in terms of mean squared error (MSE) of and R-squared (R2) metrics with MSE of 0.024 and R2 of 0.274, underscoring its effectiveness in predicting food security levels while the MLPRegressor demonstrates worst performance with MSE of 670.742 and R2 of -20636.457. The ensemble stacking model from the ensemble emerged as the best overall performer with MSE of 0.022 and R2 of 0.312, surpassing individual models and highlighting the significance of ensemble and multimodal learning approaches. These insights contribute to advancing precision agriculture and fostering resilience in farming communities, with implications for sustainable development and improved food security in Uganda. Furthermore, this study incorporates eXplainable Artificial Intelligence (XAI) techniques to interpret and visualize model predictions, providing farmers with transparent insights into the factors and key features influencing agricultural outcomes. XAI methods such as SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) are employed to enhance the transparency and trustworthiness of machine learning models, empowering farmers with actionable insights for sustainable farming practices.
- 8 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 271-281
Open AccessTransparent Multi-modelling for Climatic Geo-intelligence in Predicting Food Security
Millions of people in Uganda suffer from food insecurity as a result of the country’s scarce resources, extreme poverty, and high rate of childhood malnutrition. Women, who play a critical role in home management, are particularly vulnerable to food insecurity brought on by climate change. The vision is to foster a thriving agricultural sector in Uganda grounded in sustainability, with a mission to empower farmers, including women, through innovative tools and knowledge to adapt and prosper in a changing climate. This study employs cutting-edge technologies such as geospatial data analysis, machine learning, and environmental science to analyze weather patterns, soil conditions, and crop data. By providing personalized insights and recommendations to farmers, the research enables informed decision-making, optimizes resource management, and mitigates climate-related risks in agriculture. The research utilizes a multimodal architecture incorporating 13 traditional base machine learning models including Linear Regression, Lasso, Ridge, Elastic Net, Random Forest, K-Nearest Neighbors, Decision Tree, Support Vector Machine, Gaussian Process Regression, Gradient Boosting Regressor, XGBoost, LightGBM, and MLPRegressor. These base models are integrated into an ensemble approach using Stacking and Voting Regressors to harness collective intelligence and enhance predictive accuracy in optimizing agricultural productivity and resilience. The findings reveal that among the base models, LightGBM demonstrates superior performance in terms of mean squared error (MSE) of and R-squared (R2) metrics with MSE of 0.024 and R2 of 0.274, underscoring its effectiveness in predicting food security levels while the MLPRegressor demonstrates worst performance with MSE of 670.742 and R2 of -20636.457. The ensemble stacking model from the ensemble emerged as the best overall performer with MSE of 0.022 and R2 of 0.312, surpassing individual models and highlighting the significance of ensemble and multimodal learning approaches. These insights contribute to advancing precision agriculture and fostering resilience in farming communities, with implications for sustainable development and improved food security in Uganda. Furthermore, this study incorporates eXplainable Artificial Intelligence (XAI) techniques to interpret and visualize model predictions, providing farmers with transparent insights into the factors and key features influencing agricultural outcomes. XAI methods such as SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) are employed to enhance the transparency and trustworthiness of machine learning models, empowering farmers with actionable insights for sustainable farming practices.
- Open Access
- 8 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 451-459
Interpretable Ensemble Model-Agonistc Approaches for Malaria Prediction
Healthcare workers and patients alike face the problem of inadequate diagnosis, poor risk assessment, inadequate treatment, or toxic medication that puts patients at risk of higher death and morbidity in areas where malaria is endemic. For the past ten years, efforts to battle malaria have led to the creation of a variety of strategies, including studies on the disease that focus primarily on the automation of tools for diagnostic testing (RDT) and microscope imaging. In our work, we provide various ensemble learning techniques that provide explainability and lead to better malarial status prediction. The study focused on ensembling different individually trained models best on their accuracy performance. The results of disease prediction were made transparent and comprehensible by using explainable AI approaches like Explain Like I’m 5 (ELI5) and Local Interpretable Model-agnostic Explanation (LIME). Compared to other ensemble models, Bagging and Stacking ensemble models yielded a 99% performance. The application of these models could have a substantial influence on healthcare by decreasing the use of over-the-counter self-medication, improving patient outcomes by accurately classifying malaria cases, and increasing the effectiveness of early prediction accuracy. The explainability of the Ensemble models bolsters the relevance. Our research study enhances the field of interpretable machine learning models for medical contexts and makes a significant contribution to global public health initiatives by expanding our understanding of these models.
- 8 August 2024
- Published by Association for Computing Machinery (ACM)
- p. 451-459
Open AccessInterpretable Ensemble Model-Agonistc Approaches for Malaria Prediction
Healthcare workers and patients alike face the problem of inadequate diagnosis, poor risk assessment, inadequate treatment, or toxic medication that puts patients at risk of higher death and morbidity in areas where malaria is endemic. For the past ten years, efforts to battle malaria have led to the creation of a variety of strategies, including studies on the disease that focus primarily on the automation of tools for diagnostic testing (RDT) and microscope imaging. In our work, we provide various ensemble learning techniques that provide explainability and lead to better malarial status prediction. The study focused on ensembling different individually trained models best on their accuracy performance. The results of disease prediction were made transparent and comprehensible by using explainable AI approaches like Explain Like I’m 5 (ELI5) and Local Interpretable Model-agnostic Explanation (LIME). Compared to other ensemble models, Bagging and Stacking ensemble models yielded a 99% performance. The application of these models could have a substantial influence on healthcare by decreasing the use of over-the-counter self-medication, improving patient outcomes by accurately classifying malaria cases, and increasing the effectiveness of early prediction accuracy. The explainability of the Ensemble models bolsters the relevance. Our research study enhances the field of interpretable machine learning models for medical contexts and makes a significant contribution to global public health initiatives by expanding our understanding of these models.
- Open Access
- 24 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 163-169
A Deep Learning Approach to Detect Severity of Mango Damage in the Early Fruiting Stage
Accurate detection of mango fruit damage predominantly from fruit fly infestation is pivotal as it directly affects both yield and trade worldwide. Therefore, timely identification of such damage is critical to mitigating the spread of infestation and minimizing associated losses. This paper focuses on the early detection of mango damage in orchards using YOLOv8 models, that offer enhanced accuracy and speed compared to earlier versions, making them more efficient for object detection tasks. Limited studies have been done to detect and classify damage on fruits in orchards using deep learning, with the need for more models to detect various categories of damage instances. The experiments in this study revealed no substantial differences among the various YOLOv8 versions used with the highest accuracy of 88.6% and 98.5% attained for detecting damage and mango instances respectively. Both YOLOv8s and YOLOv8l obtained a precision value of 88.6% for lesion detection, and 87.9% using YOLOv8x. However, YOLOv8x achieved slightly higher values of recall and mAP compared to other models in detecting damage features. The study has further revealed that learning the damaged features of mango fruit is more challenging compared to healthy features, as observed from values obtained from the precision-recall curve. Through fine-tuning parameters of the models, our experimental results using the YOLOv8 model demonstrate the potential of lesion detection on mango fruits on trees, leveraging a dataset of 1317 images augmented to 3161. This study addresses the challenge of estimating profits and losses for fruits still on trees, which has been relatively overlooked in prior research efforts. We believe this method can effectively be adapted to detecting lesions on other fruits in orchards with minimal modifications. Future work can consider a better dataset with minimal noise while exploring different growth stages of fruits, and weather conditions the data is captured using alternative models while incorporating other factors in the segmentation and analysis phases.
- 24 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 163-169
Open AccessA Deep Learning Approach to Detect Severity of Mango Damage in the Early Fruiting Stage
Accurate detection of mango fruit damage predominantly from fruit fly infestation is pivotal as it directly affects both yield and trade worldwide. Therefore, timely identification of such damage is critical to mitigating the spread of infestation and minimizing associated losses. This paper focuses on the early detection of mango damage in orchards using YOLOv8 models, that offer enhanced accuracy and speed compared to earlier versions, making them more efficient for object detection tasks. Limited studies have been done to detect and classify damage on fruits in orchards using deep learning, with the need for more models to detect various categories of damage instances. The experiments in this study revealed no substantial differences among the various YOLOv8 versions used with the highest accuracy of 88.6% and 98.5% attained for detecting damage and mango instances respectively. Both YOLOv8s and YOLOv8l obtained a precision value of 88.6% for lesion detection, and 87.9% using YOLOv8x. However, YOLOv8x achieved slightly higher values of recall and mAP compared to other models in detecting damage features. The study has further revealed that learning the damaged features of mango fruit is more challenging compared to healthy features, as observed from values obtained from the precision-recall curve. Through fine-tuning parameters of the models, our experimental results using the YOLOv8 model demonstrate the potential of lesion detection on mango fruits on trees, leveraging a dataset of 1317 images augmented to 3161. This study addresses the challenge of estimating profits and losses for fruits still on trees, which has been relatively overlooked in prior research efforts. We believe this method can effectively be adapted to detecting lesions on other fruits in orchards with minimal modifications. Future work can consider a better dataset with minimal noise while exploring different growth stages of fruits, and weather conditions the data is captured using alternative models while incorporating other factors in the segmentation and analysis phases.