Estimating soil biophysical parameters from multispectral imagery is essential for precision agriculture, yet deploying accurate machine learning pipelines presents significant computational challenges. Individual soil trace elements often require specific model configurations, making traditional exhaustive hyperparameter optimization methods prohibitively time-consuming for resource-constrained embedded systems. In this study, we present a rigorous comparative analysis of two hyperparameter optimization strategies for multi-target soil estimation: the conventional, element-specific grid search, and the multi-objective Non-dominated Sorting Genetic Algorithm II (NSGA-II). Moreover, three explainable artificial intelligence methods were employed to analyze the behavior of the developed models and to provide insights into the contribution of input features to their predictions.

Algorithms developed for intelligent environments often require pattern extraction, analysis, and interpretation from multimodal and multidimensional data. This study evaluates the use of convolutional transformers for processing hyperspectral images. Since these images have a rich spatial dimension that provides valuable information, the paper demonstrates that spectral-wise methods can be used to build effective classifiers. Verification is provided using actual data from the HyBEAR dataset, which defines a binary classification task for detecting bare soil. This extensive hyperspectral image collection enabled testing several classical machine learning classifiers and comparing them with convolutional transformer-based classifiers built on the Rocket model. The verified method’s final performance yielded an F1 score of 0.904 on the emerging bare soil detection task.

Abstract. Detecting bare soil areas is an important step in the analysis of Earth observation data in a variety of Precision Agriculture (PA) applications focused on quantifying soil properties and assessing soil quality. In this paper, we introduce the HyBEAR benchmark – a novel large-scale collection of high-resolution hyperspectral aerial images (with 2 m ground sampling distance) accompanied with manual bare soil annotations verified with domain experts. Usually, the bare soil detection problem is tackled at the pixel level, meaning that detection methods classify all pixels as either bare soil or background. In contrast to this approach, we provide pixel-level annotations for the entire agricultural parcels (if the parcel is labeled as bare soil, then all pixels within that parcel are labeled accordingly), and aim to support the development of methods that identify entire fields with no vegetation. Commonly, such fields undergo further analysis to determine specific soil parameters and characteristics that are important while planning various PA activities, such as fertilization. The HyBEAR🐻 benchmark includes (i) the largest-to-date (108,064,591 pixels, corresponding to 43,225 hectares) and most heterogeneous dataset for bare soil detection, as well as (ii) the validation procedure (training-test splits and quality metrics) and a set of baseline results, obtained for a set of machine learning bare soil detection models. From the FULL collection of 1954 images in HyBEAR, which we divided into 5 spatially-disjoint folds, we additionally selected a random, stratified subset (MINI) of the images which may be useful for designing and verifying bare soil detection algorithms. Overall, HyBEAR is a step toward standardizing the way the community builds and confronts bare soil detection algorithms in a thorough, reproducible, and unbiased way.

Poland has emerged as one of the primary destinations for Ukrainian refugees, with similar to 953,000 forced migrants from Ukraine residing in the country as of May 2024. This paper examines the differences in labour market integration between Ukrainian voluntary migrants and forced migrants in Poland, comparing their situations from 2022 to 2024. The analysis is based on three quantitative surveys conducted in 2022, 2023 and 2024. The article compares the situation of both groups of migrants in several aspects: labour market situation, job satisfaction, knowledge of spoken Polish and migrants' residence plans. Our research revealed that the differences between the two groups evolved throughout the study period. Among the surveyed group of Ukrainian refugees, the proportion of those entering employment has increased significantly, reaching 64% in 2024. In the same year, the employment rate among pre-war migrants was higher, at 75.7%. The survey shows that the knowledge of the Polish language among refugees has improved over the analysed period. In the first edition of the survey, more than 20% of respondents did not speak Polish, whereas in 2024, this number had decreased to 3.2%. Our research contributes to human capital theory by emphasising that the realisation of migrants' potential-how effectively they utilise and develop their skills, education and experiences in their new country-depends not only on their qualifications but also nature of migration and shows the contrast between voluntary and forced migrants from the same country.

Detecting bare soil is a crucial component in the soil composition analysis in Earth observation images, as it allows the identification of areas of interest in the scene. It is used to exclude those regions that do not present bare soil and, therefore, may be skipped in the more intensive analysis. This step is particularly significant for on-board applications, where the hardware constraints of an edge device (an imaging satellite) must be adhered to. These constraints include computational, memory, and power consumption requirements, all of which are critical when handling large datasets in space. In this work, we tackle bare soil detection in hyperspectral images (HSIs) using lightweight machine learning models operating on not only spectral bands but also derived vegetation indices, widely adopted in the remote sensing field. Our experiments, conducted over the airborne HSIs, demonstrate that our pipeline outperforms widely established algorithms that threshold vegetation indices while delineating bare soil. It offers flexibility and fast operation, thus it may be applicable on board edge devices, such as satellites equipped with hyperspectral sensors and artificial intelligence capabilities. For reproducibility, we made the feature sets and code available at Hugging Face: https://huggingface.co/KPLabs/LightweightML-BareSoilDetection.

Satellite telemetry data consists of information transmitted from spacecraft providing their location, status, health and the functioning of the instruments. Due to the data volume, signal quality issues, resolution and orbital constraints, detecting anomalies in satellite telemetry data is challenging. One of the main challenges concerned with the existing methods is the large number of false alarms, contributing to a waste of operational resources and decreased trust in the system. In this paper, we tackle this issue and present Morphological Leave-One-Out Kernel Density Estimates (Mo-LOO-KDE)—an unsupervised anomaly detection method tailored for satellite telemetry data. It combines Extreme Value Theory, a statistical framework for modeling extremes, with computer-vision morphological operations. Here, Extreme Value Theory is used to model rare, extreme events, whereas morphological operations are effectively used for noise reduction. The unusual combination of Extreme Value Theory and morphological operations drive Mo-LOO-KDE’s low false positive rate. The proposed morphological filtering method has no formal requirement for regular sampling, so the method is applicable for irregular time series as well. Furthermore, the employment of a lightweight, unsupervised method such as Mo-LOO-KDE is advantageous due to its practical utility, as annotated and representative ground-truth datasets are expensive to generate and rarely available. Utilizing the 42-month telemetry dataset from the ESA-ADB benchmark, we perform a comparison with the SOTA approach and show that Mo-LOO-KDE outperforms other unsupervised methods, demonstrating its applicability for satellite telemetry.

Hyperspectral imaging has broad applications, particularly in Earth observation and remote sensing, due to its ability to provide large-scale, non-invasive environmental and agricultural monitoring. However, its high dimensionality and acquisition conditions pose challenges for data transfer, storage, and downstream analysis. Preprocessing algorithms are thus essential for preparing hyperspectral images for further analysis, but they rely on multiple hyperparameters, and their incorrect selection degrades the performance of machine learning models operating on the pre-processed imagery. To address this, we introduce genetic algorithms both single- and multi-objective for optimizing pre-processing in multi-target hyperspectral regression. Experiments on real-world hyperspectral data demonstrate that our approach significantly enhances the models for chlorophyll content estimation from hyperspectral images.

Transformer-based machine learning models are gaining in popularity across a range of applications. In this investigation, we explore the potential of this architecture for time-series-based forecasting, with a particular focus on assessing the benefits of permuting multiple datasets during the training process. Our research utilizes a collection of 5 different datasets collected from electricity trading markets. We selected three datasets that were tested and verified that we could substantially improve the performance of the prediction model. The mean absolute errors for all three markets decreased by a minimum of 12.5%, shifting from 6.4C, 5.4C, and 3.1$ for the models based on a single dataset for the Belgian, French, and American markets to 4.2C, 4.7C and 2.7$ for models established using three datasets, respectively. Furthermore, for all three markets tested, the final transformer configuration resulted in superior metrics in comparison with the prepared baselines, exhibiting enhanced MAE, RMSE, and SMAPE scores when evaluated against Naıve and Vanilla models.

Gray-box optimization employs Walsh decomposition to obtain non-linear variable dependencies and utilize them to propose masks of variables that have a joint non-linear influence on fitness value. These masks significantly improve the effectiveness of variation operators. In some problems, all variables are non-linearly dependent, making the aforementioned masks useless. We analyze the features of the real-world instances of such problems and show that many of their dependencies may have noise-like origins. Such noise-caused dependencies are irrelevant to the optimization process and can be ignored. To identify them, we propose extending the use of Walsh decomposition by measuring variable dependency strength that allows the construction of the weighted dynamic Variable Interaction Graph (wdVIG). wdVIGs adjust the dependency strength to mixed individuals. They allow the filtering of irrelevant dependencies and re-enable using dependency-based masks by variation operators. We verify the wdVIG potential on a large benchmark suite. For problems with noise, the wdVIG masks can improve the optimizer's effectiveness. If all dependencies are relevant for the optimization, i.e., the problem is not noised, the influence of wdVIG masks is similar to that of state-of-the-art structures of this kind.

Detecting anomalous events in satellite telemetry is a critical task in space operations. It is time-consuming, error-prone and human dependent, thus automated data-driven algorithms have been emerging at a steady pace. However, there are no available datasets of real satellite telemetry with annotations to verify anomaly detection models. We address this gap and introduce the AI-ready benchmark dataset (OPSSAT-AD) containing the telemetries acquired on board OPS-SAT—a CubeSat mission, operated by the European Space Agency. The dataset is accompanied with the baseline results obtained using 30 supervised and unsupervised classic and deep machine learning algorithms. They were evaluated using the training-test dataset split introduced in this work, and we suggest a set of quality metrics which should be calculated to confront the new algorithms for anomaly detection while exploiting OPSSAT-AD. We believe that this work may become an important step toward building a fair, reproducible and objective validation procedure that can be used to quantify the capabilities of the emerging techniques in an unbiased and fully transparent way.

Migration processes have emerged as crucial social, political and economic concerns, affecting societies, industries and organisations. The challenge lies in effectively utilizing immigrants' resources. This research aims to determine how AI tools can support matching migrants' skills with labour markets in host countries. We propose the application of an ensemble learning methodology. To validate this approach, we collect data to assess the career trajectories of 248 tertiary-educated Ukrainian immigrants in Poland, a new immigration destination. Various machine learning models are evaluated using the decision tree algorithm on these feature sets. To ensure credible results, a 10-fold cross-validation procedure is employed for each training process of every submodel. This research introduces an original ensemble machine learning classifier that combines pre-selected models with the highest performance, thereby reducing the number of parameters to be investigated. Its application in determining the career paths of highly skilled migrants, specifically Ukrainians, is novel. The study offers significant implications for Central Europe, notably Poland, where migration patterns and the integration of highly skilled migrants, mainly from Ukraine, are increasingly important. Implications for Central European audience: The ensemble machine learning classifier developed in this study could aid in optimising the career paths of these migrants, combating brain waste and facilitating their successful integration into the labour market. Integrating tools like these into decision-making processes may enhance career management and contribute to Central Europe's social and economic growth.

Machine learning has vast potential to improve anomaly detection in satellite telemetry which is a crucial task for spacecraft operations. This potential is currently hampered by a lack of comprehensible benchmarks for multivariate time series anomaly detection, especially for the challenging case of satellite telemetry. The European Space Agency Benchmark for Anomaly Detection in Satellite Telemetry (ESA-ADB) aims to address this challenge and establish a new standard in the domain. It is a result of close cooperation between spacecraft operations engineers from the European Space Agency (ESA) and machine learning experts. The newly introduced ESA Anomalies Dataset contains annotated real-life telemetry from three different ESA missions, out of which two are included in ESA-ADB. Results of typical anomaly detection algorithms assessed in our novel hierarchical evaluation pipeline show that new approaches are necessary to address operators' needs. All elements of ESA-ADB are publicly available to ensure its full reproducibility.

Diagnosing plant diseases is a difficult task, but it could be made easier with the use of advanced instrumentation and the latest machine learning techniques. This paper is a further development of a previous authors study by the authors, which has been extended to provide the classification method for tomato diseases and to indicate the spectral ranges of greatest importance for this process. As tomatoes are one of the most popular and consumed vegetables, and diseases of this crop even reduce yields by up to 80 F_1 score) of 0.896 was obtained for the logistic regression based model which was evaluated on Septoria Leaf Spot disease records.

Enhancing agricultural methods through the utilization of Earth observation and artificial intelligence (AI) has emerged as a significant concern. The ability to quantify soil parameters on a large scale can play a pivotal role in optimizing the fertilization process. While techniques for noninvasive estimation of soil parameters from hyperspectral images (HSIs) exist, their validation typically occurs across different datasets and employs varying validation protocols. This diversity renders them inherently challenging (or even impossible) to compare objectively.

This study explores the neurological basis of emotions using a multi-method approach, analysing functional near-infrared spectroscopy (fNIRS) data obtained from a 26-channel device. Author’s primary objective was to examine the brain’s response to tasks that provoke emotional, imagery and affective reactions through the observation of hemodynamic changes in data. By applying data processing methods and techniques, cerebral activations corresponding to different emotional states were shown. This research enriches comprehension of how emotion and imagery tasks are processed by the brain and offers insights into aspects of brain activity during emotionally charged engagements.

The process of forecasting using temporal convolutional neural networks on multivariate market data is an established research topic due to its practical applications. Although there exists a machine learning model for this task, their performance is still limited when they are deployed for highly volatile markets. In this study, we tackle this challenge through exploiting an additional module within a machine learning model that provides the auxiliary information captured from the input data and reflecting anomalous events observed in such markets. To further enhance the capabilities of the model, its hyperparameters are optimized using the covariance matrix adaptation evolution strategy. The experiments performed on the real-world electricity price quotes for the day-ahead market showed the superiority of our approach which achieved the prediction error of sMAPE of 8.19 and rMAE of 0.319, outperforming the vanilla model, which offered the sMAPE and rMAE of 17.80 and 0.693, and the baseline TCN of 16.34 and 0.603 accordingly.

Monitoring crop irrigation levels and the soil moisture in the root zone is of paramount importance in precision agriculture (PA), as it allows practitioners to optimize the water supply. This can, in turn, directly lead to significant water savings while maintaining appropriate cultivation practices. Currently adopted in-field methods to monitor root-zone soil moisture are costly, human-dependent, and unscalable to large spatial areas. We tackle this issue and propose MoleNose---an approach for estimating root-zone soil moisture from hyperspectral images using deep learning models enhanced through particle swarm optimization. Our experimental study, performed over a real-world dataset of hyperspectral images collocated with in-situ soil moisture measurements, revealed that MoleNose significantly outperforms hand-crafted deep learning architectures in various temporal analysis scenarios, in which we estimate the root-zone soil moisture in different time points in the season.

Maintaining proper water supply is crucial for agriculture and efficient plant cultivation. In-situ soil moisture measurements lack scalability, necessitating non-invasive systems to estimate soil moisture from remotely-sensed hyperspectral images (HSIs). We propose a root-zone soil moisture estimation system using deep architectures on HSIs acquired via unmanned aerial vehicles. Experiments conducted on a meticulously prepared dataset of in-situ measurements and HSIs, collected across multiple agronomic seasons with diverse plant varieties, soil profiles, and watering scenarios, revealed that our approach achieved a mean absolute error, mean squared error, and R2 of 0.60, 0.64, and 0.80, outperforming sensor-based methods.

Satellite imaging brings exciting opportunities in an array of fields, with precision agriculture being a notable example. Soil analysis at scale with the use of Earth observation satellites coupled with on-board and on-the-ground artificial intelligence algorithms offers actionable items that may be exploited by practitioners to optimize their operations, including the fertilization process. Here, bare soil detection is a pivotal step in the processing chain to limit the detailed analysis to the areas of interest. In this paper, we tackle this task with quantum-kernel support vector machines and verify the utility of quantum machine learning in practical Earth observation. Our experimental study, performed over a real-world hyperspectral scene, indicates that the proposed quantum-kernel models are competitive with well-established classical support vector machines, as well as with approaches based on thresholding spectral indices that are widely exploited in the field.

Bare soil detection is an important step in soil composition analysis, as it can prune the areas that should be excluded from more expensive processing aimed at extracting selected soil parameters from hyperspectral images acquired in orbit. This is of paramount importance for on-board applications, where hardware constraints of an edge device (a satellite), such as computational and memory requirements or energy consumption need to be considered while processing big data in space. In this paper, we present a simple yet effective bare soil detection algorithm exploiting vegetation indices that is ready for in-orbit deployment. Our experimental study performed over the airborne hyperspectral data shows that this approach can be robustly used for simulated bands, i.e., wide bands aggregating several narrow neighboring bands within the spectrum. Therefore, we can apply our technique to sensors with lower spectral resolution. Finally, it offers high-quality bare soil delineation reaching the Dice Index of 0.85.