Marine litter pollution is a major environmental threat due to the widespread presence of plastics and their detrimental impact on marine life and human health. There is a need for autonomous systems with computer vision to help clean the oceans. This study compares the latest state-of-the-art You Only Look Once (YOLO) models YOLOv9 - YOLOv12 in an underwater object detection setting in terms of accuracy, computational speed, and architecture complexity. We specifically focus on the smallest versions of these architectures, due to the real-time constraints of the setting. Multiple underwater datasets are combined to obtain a wide representation of underwater conditions and marine objects. The findings provide valuable insights into selecting and optimizing object detection architectures for underwater litter detection, contributing to monitoring marine ecosystems and addressing marine pollution. This work can be used as a building ground for further improving underwater object detection systems.

This paper investigates the significance of hyperparameter optimization in meta-learning for image classification tasks. Despite advancements in deep learning, real-time image classification applications often suffer from data inadequacy. Few-shot learning addresses this challenge by enabling learning from limited samples. Meta-learning, a prominent tool for few-shot learning, learns across multiple classification tasks. We explore different types of meta-learners, with a particular focus on metric-based models. We analyze the potential of hyperparameter optimization techniques, specifically Bayesian optimization and its variants, to enhance the performance of these models. Experimental results on the Omniglot and ImageNet datasets demonstrate that incorporating Bayesian optimization, particularly its evolutionary strategy variant, into meta-learning frameworks leads to improved accuracy compared to settings without hyperparameter optimization. Here, we show that by optimizing hyperparameters for individual tasks rather than using a uniform setting, we achieve notable gains in model performance, underscoring the importance of tailored hyperparameter configurations in meta-learning.

Flooding is one of the most common natural hazards that have extremely detrimental consequences. Understanding which areas are vulnerable to flooding is crucial to addressing these effects. In this work, we use machine learning models and Automated machine learning (AutoML) systems for flood susceptibility mapping in Kerala, India. In particular, we used a three-dimensional convolutional neural network (CNN) architecture for this purpose. The CNN model was assisted with hyperparameter optimization techniques that combine Bayesian optimization with evolutionary algorithms like differential evolution and covariance matrix adaptation evolutionary strategies. The performances of all models are compared in terms of cross-entropy loss, accuracy, precision, recall, area under the curve (AUC) and kappa score. The CNN model shows better performance than the AutoML models. Evolutionary algorithm-assisted hyperparameter optimization methods improved the efficiency of the CNN model by 4 and 9 percent in terms of accuracy and by 0.0265 and 0.0497 with reference to the AUC score.

For any machine learning model, finding the optimal hyperparameter setting has a direct and significant impact on the model’s performance. In this paper, we discuss different types of hyperparameter optimization techniques. We compare the performance of some of the hyperparameter optimization techniques on image classification datasets with the help of AutoML models. In particular, the paper studies Bayesian optimization in depth and proposes the use of genetic algorithm, differential evolution and covariance matrix adaptation—evolutionary strategy for acquisition function optimization. Moreover, we compare these variants of Bayesian optimization with conventional Bayesian optimization and observe that the use of covariance matrix adaptation—evolutionary strategy and differential evolution improves the performance of standard Bayesian optimization. We also notice that Bayesian optimization tends to perform poorly when genetic algorithm is used for acquisition function optimization.

Multi-Model Ensembles (MMEs) are used for improving the performance of GCM simulations. This study evaluates the performance of MMEs of precipitation, maximum temperature and minimum temperature over a tropical river basin in India developed by various techniques like arithmetic mean, Multiple Linear Regression (MLR), Support Vector Machine (SVM), Extra Tree Regressor (ETR), Random Forest (RF) and long short-term memory (LSTM). The 21 General Circulation Models (GCMs) from National Aeronautics Space Administration (NASA) Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) dataset and 13 GCMs of Coupled Model Inter-comparison Project, Phase 6 (CMIP6) are used for this purpose. The results of the study reveal that the application of a LSTM model for ensembling performs significantly better than models in the case of precipitation with a coefficient of determination (R ²) value of 0.9. In case of temperature, all the machine learning (ML) methods showed equally good performance, with RF and LSTM performing consistently well in all the cases of temperature with R ² value ranging from 0.82 to 0.93. Hence, based on this study RF and LSTM methods are recommended for creation of MMEs in the basin. In general, all ML approaches performed better than mean ensemble approach.