Scientometrics is undergoing a methodological transformation driven by the increasing availability of large-scale scientific data and advances in artificial intelligence (AI). Traditional approaches centered on citation analysis and bibliographic coupling are now complemented by methods that leverage semantic representations, structured knowledge, and natural language understanding. However, current generative AI systems pose inherent challenges such as hallucinations, lack of transparency in decision-making and explainability, and issues with source reliability. In this article, we seek to mitigate these challenges by outlining a framework that integrates knowledge graphs (KGs) and large language models (LLMs) for scientometric inquiry. Taking a socio-technical perspective, the article sets out to explore the intersectional space of computing and information science from a human-centered approach. The framework is designed to support both established scientometric tasks, such as trend analysis, topic detection, and collaboration mapping, and more exploratory, insight-generating applications, including question answering, knowledge discovery, and contextual enrichment of scientific content. Drawing on recent developments in the use of KGs and LLMs in scientific domains, we provide a comparative overview of existing work, identify key design principles, and discuss the advantages and limitations of such a framework. Our goal is to chart a pathway toward more interactive, transparent, and generative approaches in information science and cross-disciplinary research. • Multidimensional mapping from large bibliographic datasets remains challenging. • A KG-LLM framework is proposed to support interactive scientometric inquiry. • HITL validation addresses the black-box limitations of KG- and LLM-enabled solutions. • Implications for knowledge-based question answering are derived.

Continual learning aims to enable models to acquire new knowledge while retaining previously learned information. Prompt-based methods have shown remarkable performance in this domain; however, they typically rely on key-value pairing, which can introduce inter-task interference and hinder scalability. To overcome these limitations, we propose a novel approach employing task-specific Prompt-Prototype (ProP), thereby eliminating the need for key-value pairs. In our method, task-specific prompts facilitate more effective feature learning for the current task, while corresponding prototypes capture the representative features of the input. During inference, predictions are generated by binding each task-specific prompt with its associated prototype. Additionally, we introduce regularization constraints during prompt initialization to penalize excessively large values, thereby enhancing stability. Experiments on several widely used datasets demonstrate the effectiveness of the proposed method. In contrast to mainstream prompt-based approaches, our framework removes the dependency on key-value pairs, offering a fresh perspective for future continual learning research.

The molecular classification of adult-type diffuse gliomas is essential for determining appropriate therapeutic strategies, but genomic sequencing remains costly. Recent advances in digital pathology and deep learning have led to several studies exploring molecular classification using multiple instance learning (MIL) on whole slide images (WSIs). However, achieving optimal classification performance using only histological slides is challenging due to the lack of guidance from genomic data. In this study, we propose a teacher-student distillation framework for glioma molecular classification using WSIs. Our method leverages a pretrained self-normalizing neural network (SNN) as the genomic teacher model, which selects genes based on survival analysis-driven criteria to guide the MIL-based student model in learning effective histological representations. During training, both genomic and pathological data are utilized, while inference relies solely on WSIs. Experimental validation on the TCGA GBM-LGG datasets shows that our approach outperforms state-of-the-art (SOTA) MIL models, highlighting its effectiveness in glioma diagnostic subtyping using WSIs.

Brain-inspired Spiking Neural Networks (SNNs) have garnered significant attention due to their bio-plausibility and low power consumption advantages compared to Artificial Neural Networks (ANNs). However, the application of SNN in computer vision remains limited, primarily due to their inferior performance. In this work, we aim to bridge the performance gap between ANNs and SNNs in object detection by our Advanced SpikingYOLOX. The proposed approach extends the SpikingYOLOX with two key innovations: PSA-SNN and 2D-Spiking Transformer, both designed to enhance object detection performance. PSA-SNN extends spike-based self-attention by incorporating high-speed partial self-attention with an SNN-based 2D-Spiking Transformer in the deepest layer of the backbone, significantly improving feature extraction. The 2D-Spiking Transformer redefines the role of spiking neurons in Transformer sequences (Key, Query, Value), demonstrating that applying an additional spiking layer solely to the Value sequence yields the best performance while maintaining computational efficiency in spike-driven Transformers. We conduct extensive experiments on static images and the Advanced SpikingYOLOX achieves state-of-the-art performance among other SNN-based object detection methods. This work paves the way for more advanced SNN applications in object detection and broader computer vision tasks.

Abstract Distance-based supervised method, the minimal learning machine, constructs a predictive model from data by learning a mapping between input and output distance matrices. In this paper, we propose new methods and evaluate how their core component, the distance mapping, can be adapted to multi-label learning. The proposed approach is based on combining the distance mapping with an inverse distance weighting. Although the proposal is one of the simplest methods in the multi-label learning literature, it achieves state-of-the-art performance for small to moderate-sized multi-label learning problems. In addition to its simplicity, the proposed method is fully deterministic: Its hyper-parameter can be selected via ranking loss-based statistic which has a closed form, thus avoiding conventional cross-validation-based hyper-parameter tuning. In addition, due to its simple linear distance mapping-based construction, we demonstrate that the proposed method can assess the uncertainty of the predictions for multi-label classification, which is a valuable capability for data-centric machine learning pipelines.

OBJECTIVE: Intracranial electroencephalogram (iEEG) signals are generally recorded using multiple channels, and channel selection is therefore a significant means in studying iEEG-based seizure prediction. For n channels, [Formula: see text] channel cases can be generated for selection. However, by this means, an increase in n can cause an exponential increase in computational consumption, which may result in a failure of channel selection when n is too large. Hence, it is necessary to explore reasonable channel selection strategies under the premise of controlling computational consumption and ensuring high classification accuracy. Given this, we propose a novel method of channel reordering strategy combined with dual CNN-LSTM for effectively predicting seizures. METHOD: First, for each patient with n channels, interictal and preictal iEEG samples from each single channel are input into the CNN-LSTM model for classification. Then, the F1-score of each single channel is calculated, and the channels are reordered in descending order according to the size of F1-scores (channel reordering strategy). Next, iEEG signals with an increasing number of channels are successively fed into the CNN-LSTM model for classification again. Finally, according to the classification results from n channel cases, the channel case with the highest classification rate is selected. RESULTS: Our method is evaluated on the three iEEG datasets: the Freiburg, the SWEC-ETHZ and the American Epilepsy Society Seizure Prediction Challenge (AES-SPC). At the event-based level, the sensitivities of 100%, 100% and 90.5%, and the false prediction rates (FPRs) of 0.10/h, 0/h and 0.47/h, are achieved for the three datasets, respectively. Moreover, compared to an unspecific random predictor, our method also shows a better performance for all patients and dogs from the three datasets. At the segment-based level, the sensitivities-specificities-accuracies-AUCs of 88.1%-94.0%-93.5%-0.9101, 99.1%-99.7%-99.6%-0.9935, and 69.2%-79.9%-78.2%-0.7373, are attained for the three datasets, respectively. CONCLUSION: Our method can effectively predict seizures and address the challenge of an excessive number of channels during channel selection.

Multi-label classification problems, where more than one class can be active in a single instance, generalize the conventional single-label cases.In this article, we continue the research track documented in [1,2], where the Minimal Learning Machine (MLM) was generalized into multilabel problems with competitive results compared to other state-of-the-art techniques.Our current interest is to consider whether we can reduce the complexity of the distance-based regression model in the MLM by performing feature selection.For this purpose, an existing feature selection filter technique is generalized to multi-label problems.Experimental results confirm that the proposed technique provides a useful ranking, which allows one to reduce the number of active features without jeopardizing the quality of the multi-label MLM classifier.

The pervasiveness of technical systems in our lives calls for a broad understanding of the interaction between humans and technology. Affinity for technology interaction (ATI) scale measures the tendency of a person to actively engage or to avoid interaction with technological systems, including both software and physical devices. This research presents a psychometric analysis of a Finnish version of the ATI scale. The data consisted of 796 responses of students in a Finnish university. The data were analyzed utilizing factor analysis and both nonparametric and parametric item response theory. The Finnish version of the ATI scale proved to be essentially unidimensional, showing high reliability estimates, and forming a strong Mokken scale. Hierarchical multiple regression analysis showed that men had a slightly higher affinity for technology than women when controlling for age and field of study; however the effect size was small.

Dimension reduction is one of the key data transformation techniques in machine learning and knowledge discovery. It can be realized by using linear and nonlinear transformation techniques. An additive autoencoder for dimension reduction, which is composed of a serially performed bias estimation, linear trend estimation, and nonlinear residual estimation, is proposed and analyzed. Compared to the classical model, adding an explicit linear operator to the overall transformation and considering the nonlinear residual estimation in the original data dimension significantly improves the data reproduction capabilities of the proposed model. The computational experiments confirm that an autoencoder of this form, with only a shallow network to encapsulate the nonlinear behavior, is able to identify an intrinsic dimension of a dataset with low autoencoding error. This observation leads to an investigation in which shallow and deep network structures, and how they are trained, are compared. We conclude that the deeper network structures obtain lower autoencoding errors during the identification of the intrinsic dimension. However, the detected dimension does not change compared to a shallow network. As far as we know, this is the first experimental result concluding no benefit from a deep architecture compared to its shallow counterpart.

Feature selection (FS) may improve the performance, cost-efficiency, and understandability of supervised machine learning models. In this paper, FS for the recently introduced distance-based supervised machine learning model is considered for regression problems. The study is contextualized by first providing an umbrella review (review of reviews) of recent development in the research field. We then propose a saliency-based one-shot wrapper algorithm for FS, which is called MAS-FS. The algorithm is compared with a set of other popular FS algorithms, using a versatile set of simulated and benchmark datasets. Finally, experimental results underline the usefulness of FS for regression, confirming the utility of certain filter algorithms and particularly the proposed wrapper algorithm.

Interest in multi-target regression and multi-label classification techniques and their applications have been increasing lately. Here, we use the distance-based supervised method, minimal learning machine (MLM), as a base model for multi-label classification. We also propose and test a hybridization of unsupervised and supervised techniques, where prototype-based clustering is used to reduce both the training time and the overall model complexity. In computational experiments, competitive or improved quality of the obtained models compared to the state-of-the-art techniques was observed.

Machine learning (ML) force fields are one of the most common applications of ML in nanoscience.However, commonly these methods are trained on potential energies of atomic systems and force vectors are omitted.Here we present a ML framework, which tackles the greatest difficulty on using forces in ML: accurate prediction of force direction.We use the idea of Minimal Learning Machine to device a method which can adapt to the orientation of an atomic environment to estimate the directions of force vectors.The method was tested with linear alkane molecules.

This paper presents a novel approach and a method of learning analytics to study student agency in higher education. Agency is a concept that holistically depicts important constituents of intentional, purposeful, and meaningful learning. Within workplace learning research, agency is seen at the core of expertise. However, in the higher education field, agency is an empirically less studied phenomenon with also lacking coherent conceptual base. Furthermore, tools for students and teachers need to be developed to support learners in their agency construction. We study student agency as a multidimensional phenomenon centring on student-experienced resources of their agency. We call the analytics process developed here student agency analytics, referring to the application of learning analytics methods for data on student agency collected using a validated instrument. The data are analysed with unsupervised and supervised methods. The whole analytics process will be automated using microservice architecture. We provide empirical characterisations of student-perceived agency resources by applying the analytics process in two university courses. Finally, we discuss the possibilities of using agency analytics in supporting students to recognise their resources for agentic learning and consider contributions of agency analytics to improve academic advising and teachers' pedagogical knowledge.

Epileptic seizure detection using scalp electroencephalogram (sEEG) and intracranial electroencephalogram (iEEG) has attracted widespread attention in recent two decades. The accurate and rapid detection of seizures not only reflects the efficiency of the algorithm, but also greatly reduces the burden of manual detection during long-term electroencephalogram (EEG) recording. In this work, a stacked one-dimensional convolutional neural network (1D-CNN) model combined with a random selection and data augmentation (RS-DA) strategy is proposed for seizure onset detection. Firstly, we segmented the long-term EEG signals using 2-s sliding windows. Then, the 2-s interictal and ictal segments were classified by the stacked 1D-CNN model. During model training, a RS-DA strategy was applied to solve the problem of sample imbalance, and the patient-specific model was trained with event-based K-fold (K is the number of seizures per patient) cross validation for detecting all seizures of each patient. Finally, we evaluated the performances of the proposed approach in the two levels: the segment-based level and the event-based level. The proposed method was tested on two long-term EEG datasets: the CHB-MIT sEEG dataset and the SWEC-ETHZ iEEG dataset. For the CHB-MIT sEEG dataset, we achieved 88.14% sensitivity, 99.62% specificity and 99.54% accuracy in the segment-based level. From the perspective of the event-based level, 99.31% sensitivity, 0.2/h false detection rate (FDR) and mean 8.1-s latency were achieved. For the SWEC-ETHZ iEEG dataset, in the segment-based level, 90.09% sensitivity, 99.81% specificity and 99.73% accuracy were obtained. In the event-based level, 97.52% sensitivity, 0.07/h FDR and mean 13.2-s latency were attained. From these results, we can see that our method can effectively use both sEEG and iEEG data to detect epileptic seizures, and this may provide a reference for the clinical application of seizure onset detection.

This Research Full Paper presents an examination of the relationships between course satisfaction and student agency resources in engineering education. Satisfaction experienced in learning is known to benefit the students in many ways. However, the varying significance of the different factors of course satisfaction is not entirely clear. We used a validated questionnaire instrument, exploratory statistics, and supervised machine learning to examine how the different factors of student agency affect course satisfaction among engineering students (N = 293). Teacher's support and trust for the teacher were identified as both important and critical factors concerning experienced course satisfaction. Participatory resources of agency and gender proved to be less important factors. The results provide convincing evidence about the possibility to identify the most important factors affecting course satisfaction.

In the recent years, multilevel inverter topologies have gained a lot of attention for their ability to produce improved power quality. This paper addresses the operation of a nine-level single-phase inverter called Packed E-Cell Converter. The effects of dead time on the output voltage waveform is under study for the inverter topology. It is shown, that the topology lacks freewheeling paths during some of the commutations that would provide smooth transitions among the nine output voltage levels. With the original modulation, even a full DC link voltage transient pulse may occur during a dead time, which is four times as large as the intended voltage step. In this paper, an alternative switching state logic is proposed to improve the inverter output voltage waveform and the produced common-mode voltage. The original and proposed switching state schemes are simulated to demonstrate the obstacles to the applicability of the inverter topology.

Internet of things (IoT) and digitalization of industries are changing the skills required from newly graduated engineers. To truly master these skills, they must be included in the curriculum, and a platform on which students can create IoT projects is needed. The research of IoT and related areas also benefits from such a platform. The platform presented in this paper focuses on a simple coffee maker but applies techniques and tools which are also used in larger scale projects, thereby making the system a valid solution for education. The devices and tools applied must be selected to satisfy education and research requirements. The proposed platform is built on IBM Cloud, and it uses Raspberry Pi as an edge device. Hands-on experience with a real IoT platform brings definite advantages that may otherwise be difficult to achieve.

Industrial automation systems have collected vast amounts of data for years. Data analytics and machine learning can be used to reveal different phenomena and anomalies, which may be otherwise impossible to see. However, the opportunities offered by the data are not currently utilized even though the technology is available. In this paper, a the potential use of the data analytics and machine learning of automation system data is presented. A case study on indirect measurement and predictive analysis of electric motor overcurrent was carried out in a pulp mill. Predictive models reached accuracy up to 98,85 %. The methods presented can be generalized to other processes. Since automation systems store data in most industrial sites, no additional hardware is necessarily needed for industrial internet of things (IIoT) systems, making a factory scale IIoT system possible.

Short-term outreach interventions are conducted to raise young students’ awareness of the computer science (CS) field. Typically, these interventions are targeted at K–12 students, attempting to encourage them to study CS in higher education. This study is based on a series of extra-curricular outreach events that introduced students to the discipline of computing, nurturing creative computational thinking through problem solving and game programming. To assess the long-term impact of this campaign, the participants were contacted and interviewed two to five years after they had attended an outreach event. We studied how participating in the outreach program affected the students’ perceptions of CS as a field and, more importantly, how it affected their educational choices. We found that the outreach program generally had a positive effect on the students’ educational choices. The most prominent finding was that students who already possessed a “maintained situational interest” in CS found that the event strengthened their confidence in studying CS. However, many students were not affected by attending the program, but their perceptions of CS did change. Our results emphasize the need to provide continuing possibilities for interested students to experiment with computing-related activities and hence maintain their emerging individual interests.

The purpose of academic advising is to help students with developing educational plans that support their academic career and personal goals, and to provide information and guidance on studies. Planning and management of the students’ study path is the main joint activity in advising. Based on a study log of passed courses, we propose to use robust, prototype-based clustering to identify a set of actual study path profiles. Such profiles identify groups of students with similar progress of studies, whose analysis and interpretation can be used for better institutional awareness and to support evidence-based academic advising. A model of automated academic advising system utilizing the possibility to determine the study profiles is proposed.

Online changepoint detection is an important task for machine learning in changing environments, as it signals when the learning model needs to be updated. Presence of noise that can be mistaken for real changes makes it difficult to develop an effective approach that would have a low false alarm rate and being able to detect all the changes with a minimal delay. In this paper we study how performance of popular Bayesian online detectors can be improved in case of recurrent changes. Modelling recurrence allows us to anticipate future changepoints and predict their locations in time. We propose an approach for inducing and integrating recurrence information in the streaming settings, and demonstrate its effectiveness on synthetic and realworld human activity datasets.

The task of online change point detection in sensor data streams is often complicated due to presence of noise that can be mistaken for real changes and therefore affecting performance of change detectors. Most of the existing change detection methods assume that changes are independent from each other and occur at random in time. In this paper we study how performance of detectors can be improved in case of recurrent changes. We analytically demonstrate under which conditions and for how long recurrence information is useful for improving the detection accuracy. We propose a simple computationally efficient message passing procedure for calculating a predictive probability distribution of change occurrence in the future. We demonstrate two straightforward ways to apply the proposed procedure to existing change detection algorithms. Our experimental analysis illustrates the effectiveness of these approaches in improving the performance of a baseline online change detector by incorporating recurrence information.

Cyber resilient organizations, their functions and computing infrastructures, should be tolerant towards rapid and unexpected changes in the environment. Information security is an organization-wide common mission; whose success strongly depends on efficient knowledge sharing. For this purpose, semantic wikis have proved their strength as a flexible collaboration and knowledge sharing platforms. However, there has not been notable academic research on how semantic wikis could be used as information security management platform in organizations for improved cyber resilience. In this paper, we propose to use semantic wiki as an agile information security management platform. More precisely, the wiki contents are based on the structured model of the NIST Special Publication 800-53 information security control catalogue that is extended in the research with the additional properties that support the information security management and especially the security control implementation. We present common uses cases to manage the information security in organizations and how the use cases can be implemented using the semantic wiki platform. As organizations seek cyber resilience, where focus is in the availability of cyber-related assets and services, we extend the control selection with option to focus on availability. The results of the study show that a semantic wiki based information security management and collaboration platform can provide a cost-efficient solution for improved cyber resilience, especially for small and medium sized organizations that struggle to develop information security with the limited resources.

Variational methods are commonly used to solve noise removal problems. In this paper, we present an augmented Lagrangian-based approach that uses a discrete form of the $L^1$-norm of the mean curvature of the graph of the image as a regularizer, discretization being achieved via a finite element method. When a particular alternating direction method of multipliers is applied to the solution of the resulting saddle-point problem, this solution reduces to an iterative sequential solution of four subproblems. These subproblems are solved using Newton’s method, the conjugate gradient method, and a partial solution variant of the cyclic reduction method. The approach considered here differs from existing augmented Lagrangian approaches for the solution of the same problem; indeed, the augmented Lagrangian functional we use here contains three Lagrange multipliers “only,” and the associated augmentation terms are all quadratic. In addition to the description of the solution algorithm, this paper contains the results of numerical experiments demonstrating the performance of the novel method discussed here.

When we introduce icon-based language into the context of requirements engineering, we must take into account that what users perceive as recognizable and usable depends on their background. In this paper, we argue that it is not possible to provide a single set of visual notations that appeal to all of stakeholders. Instead, we suggest an adaptable preference framework, which generates personalized notations that correspond to personal background. We present and evaluate icon-based language: a new kind of approach to requirements engineering work to explore its possibility and usability. In an initial evaluation of students residing in Finland, results reveal that users are able to recognize a group of icons fairly well. Our findings show that an icon-based language could probably be a positive means in improving awareness of requirements engineering as it tends to take advantages of icons which are intuitively understandable to represent traditional textual requirements.

Data masses require a lot of data processing. Data mining is the traditional way to convert data into knowledge. In visual analytics, humans are integrated into the process as there is continuous interaction between the analyst and the analysis software. Data mining methods can be utilized also in visual analytics where the priority is given to the visualization of the information and to dimension reduction. However, the provided data is not always enough. There is a large amount of background contextual information, which should be included into the automated process. This paper describes a context-sensitive approach, in which we utilize visual analytics by studying all phases in the process according to our “sensing, processing and actuation” framework. Experimental studies show that our framework can be very useful in the process of analyzing causes for and relations between variable changes with laboratory-scale power plant data.

This paper considers the difficulties faced by the stakeholders in general requirements engineering (RE). These difficulties range from the complexity of requirements gathering to requirements presentation. Affordable visualization techniques have been widely implemented to support the requirements engineering community. However, no universal characteristics that could be associated with requirements completion have been identified so far. The research focus of this paper is driven by the above considerations to introduce the icon-based language comprising a set of icon notations, syntactic and semantics. Icon-based language would support the requirements engineering tasks that normally executed by stakeholders and provide a visual modelling language to unify the requirement activities. Research approach is recapitulate, firstly, by identifying the requirements engineering artefact, secondly by refining the icon artefact, and thirdly, by integrating those two artefacts by means of requirements engineering process. The result aimed at to make communications more interactive and manageable by facilitating the exchange of information and to promote global understanding in any requirements development context across cultural and national boundaries.

Data masses require a lot of data processing. Data mining is the traditional way to transfer data into knowledge. In visual analytics, humans are integrated into the process due to continuous interaction between the analyst and the analysis software. Data mining methods can be utilized also in visual analytics, where the priority is given to the visualization of the information and to dimension reduction. However, the provided data is not always enough. There is a large amount of background contextual information, which should be included into the automated process. This paper describes a context-sensitive approach, in which we utilize visual analytics by studying all phases in the process. To observe the process from all points of view, we divide it into three architectural sections: sensing, processing, and actuation. Thanks to the versatility of visual analytics methods, our goal in the future is to apply these methods in the field of energy production in a laboratory-scale power plant.

Context is a multi-dimensional concept. It is hard to define context generally for computer science. Which information is considered as context, which is not? Why are the certain context elements relevant for a certain case, but irrelevant for another? How to explain this to computers? Can computers learn these issues as humans do? In our paper we present different viewpoints to the concept of context and to context modelling starting from requirements engineering and ending up to multi-disciplinary education. Based on context related literature research and discussions in our paper, we can summarize that a complete and comprehensive definition and model of context is difficult to achieve and may not even be appropriate at all. However we can conclude that there is a common understanding that context always relates to an entity, context is used to solve a problem, context depends on the domain of use, context depends on time and context is evolutionary.

Multilayer perceptrons (MLP) are often trained by minimizing the mean of squared errors (MSE), which is a sum of squared Euclidean norms of error vectors. Less common is to minimize the sum of Euclidean norms without squaring them. The latter approach, mean of non-squared errors (ME), bears implications from robust statistics. We carried out computational experiments to see if it would be notably better to train an MLP classifier by minimizing ME instead of MSE in the special case when training data contains class noise, i.e., when there is some mislabeling. Based on our experiments, we conclude that for small datasets containing class noise, ME could indeed be a very preferable choice, whereas for larger datasets it may not help.

This paper is the part II of a paper composed of two parts. In the part I, a memetic approach consisting of applying a local search to the scale factor of a differential evolution framework in order to generate an off-spring with a high quality was proposed. The part II proposes the application of the scale factor local search within a differential evolution framework which integrates a self-adaptive update of the control parameters. In other words, unlike for the part I, the scale factor local search is applied to a an algorithmic framework characterized by multiple scale factors over the individuals of the population and scale factor updates during the evolution. Two simple local search logics have been tested, the first one employs the golden section search and the second one a hill-climber. The local search algorithms thus assist the global search and generates offspring with high performance which are subsequently supposed to promote the generation of better solutions within the evolutionary framework. Numerical results show that the hybridization is beneficial and able to outperform in many cases both the classical differential evolution and a self-adaptive differential evolution recently proposed in literature.

In this article, a method for separating linear subspaces of time-locked brain responses and other noise sources in multichannel electroencephalography data is proposed. The components related to time-locked and noise subspaces are distinguished by method based on different behavior they experience after traditional averaging. The actual separation of the two subspaces is performed without whitening by maximizing/minimizing the same criterion. The detailed derivation of the method is given, and the results of the method's application to simulated and real EEG datasets are studied. The possibilities of improving the results are also discussed.

This paper proposes the employment of continuous probability distributions instead of step functions for adaptive coordination of the local search in fitness diversity based memetic algorithms. Two probability distributions are considered in this study: the beta and exponential distributions. These probability distributions have been tested within two memetic frameworks present in literature. Numerical results show that employment of the probability distributions can be beneficial and improve performance of the original memetic algorithms on a set of test functions without varying the balance between the evolutionary and local search components.

We argue for the importance of a real customer in a capstone project course via reporting experiences from over ten years period at University of Jyväskylä (JYU), and providing also practical knowledge and theoretical underpinnings for such an arrangement. Based on our experience and empirical data, we underline that this arrangement is a chance for the students to increase their occupational identity. Theoretically our rationalization of the course arrangement proposes self-direction, empowerment, motivation, and deep level learning strategies for the students.

This article proposes an Enhanced Memetic Differential Evolution (EMDE) for designing digital filters which aim at detecting defects of the paper produced during an industrial process. Defect detection is handled by means of two Gabor filters and their design is performed by the EMDE. The EMDE is a novel adaptive evolutionary algorithm which combines the powerful explorative features of Differential Evolution with the exploitative features of three local search algorithms employing different pivot rules and neighborhood generating functions. These local search algorithms are the Hooke Jeeves Algorithm, a Stochastic Local Search, and Simulated Annealing. The local search algorithms are adaptively coordinated by means of a control parameter that measures fitness distribution among individuals of the population and a novel probabilistic scheme. Numerical results confirm that Differential Evolution is an efficient evolutionary framework for the image processing problem under investigation and show that the EMDE performs well. As a matter of fact, the application of the EMDE leads to a design of an efficiently tailored filter. A comparison with various popular metaheuristics proves the effectiveness of the EMDE in terms of convergence speed, stagnation prevention, and capability in detecting solutions having high performance.

This paper compares three different fitness diversity adaptations in multimeme algorithms (MmAs). These diversity indexes have been integrated within a MmA present in literature, namely fast adaptive memetic algorithm. Numerical results show that it is not possible to establish a superiority of one of these adaptive schemes over the others and choice of a proper adaptation must be made by considering features of the problem under study. More specifically, one of these adaptations outperforms the others in the presence of plateaus or limited range of variability in fitness values, another adaptation is more proper for landscapes having distant and strong basins of attraction, the third one, in spite of its mediocre average performance can occasionally lead to excellent results.

The connection between robust statistical estimates and nonsmooth optimization is established. Based on the resulting family of optimization problems, robust learning problem formulations with regularization-based control on the model complexity of the multilayer perceptron network are described and analyzed. Numerical experiments for simulated regression problems are conducted, and new strategies for determining the regularization coefficient are proposed and evaluated.

A simple and general calculus for the sensitivity analysis of a feedforward MLP network in a layer-wise form is presented. Based on the local optimality conditions, some consequences for the least-means-squares learning problem are stated and further discussed. Numerical experiments with formulation and comparison of different weight decay techniques are included.