Efficiently comprehending vast volumes of text, particularly in scientific literature, is a formidable challenge for researchers. To alleviate this burden, this paper investigates the application of topic modeling and extractive summarization techniques to analyze and distill valuable insights from scientific documents. Focusing on the NeurIPS dataset, which encompasses a diverse array of scientific papers, we employ techniques such as Latent Dirichlet Allocation (LDA) and probabilistic Latent Semantic Analysis (pLSA) to categorize documents based on their content. Subsequently, we implement a domain-specific extractive summarizer that leverages Convolutional Neural Networks (CNNs) for feature extraction and semantic segmentation. The summarizer is trained to identify and rank sentences based on relevance, using binarized ROUGE-2 scores as a metric. Our results demonstrate the effectiveness of these methods in condensing scientific documents, facilitating quicker information retrieval and understanding. The paper also compares the performance of LDA and pLSA, highlighting LDA’s superiority in topic coherence and generalizability. By leveraging these methods, we aim to demonstrate their efficacy in processing large volumes of text, facilitating enhanced information retrieval and understanding in the realm of scientific literature.

Amidst an era of speedy technological growth, fraud is a complex challenge. This article presents an innovative analytical method that uses utility consumption data to identify probable instances of fraud in utility-based services. This study evaluates the efficacy of neural networks, specifically Artificial Neural Network (ANN) — a subset of neural networks, compared to Light Gradient Boosting Machine (LGBM), a tree-based model. It uses data from the Zindi challenge to analyze their capability to detect fraudulent consumption patterns. The article focuses on optimizing features using Pearson correlation and discusses the difficulties associated with imbalanced data sets. The LGBM model has exceptional performance, as seen by its impressive ROC AUC score of 0.878242. This number highlights its remarkable ability to differentiate fraudulent actions from ANN, thus establishing future advancements in fraud detection in utility-based services.

In recent years, attention within the clinical prediction community has turned to the use of survival machine learning as a tool for predicting the risk of developing a disease as a function of time. The current work seeks to contribute to existing literature which demonstrates the utility of these methods when applied to a dementia prediction context. We use the Alzheimer's Disease Neuroimaging Initiative ADNI dataset and model deterioration within two distinct groups, those deemed cognitively normal and those with a formal diagnosis of Mild Cognitive Impairment. In agreement with existing literature we find that survival machine learning outperforms standard survival analysis methods such as Cox PH model, and has very good predictive ability. We propose an innovative approach to predicting dementia diagnosis risk on ADNI, which explores the use of survival neural network and survival extreme gradient boosting techniques that have hitherto seldom been applied to this context. The stability of our models was investigated within a Monte Carlo simulation framework.

Artificial intelligence (AI) is one of computer science's most crucial subfields. AI applications are growing rapidly. Games are the main place where artificial intelligence is actually used. AI is used to create the game's non-player characters, also referred to as NPCs. Although there are a lot of alternative ways to implement AI in a game, the search strategy is by far the most common. A variety of different search algorithms can be used to construct AI in video games. This paper's main objective is to use the snake game as a comparative tool to examine the differences between search algorithms employed by human agents and those used in AI. This paper presents detailed assessments of informed and uninformed search strategies as well as experiments on Hamiltonian search. This paper also addresses a few widely used techniques, like finite state machines, behavior trees, tree search, etc., frequently used while creating AI. We get to the conclusion that some algorithms are better candidates for use as search algorithms based on the results of thorough expla-nations and trials. We found that the Human Agent performs poorly when compared to the Depth-First Search, Breadth-First Search, Hamiltonian Search, and Best-First Search algorithms, while the A * Search algorithm surpasses them all significantly.

Dynamic optimization is of significant practical relevance for production and delivery processes that involve storage systems because of the dynamic nature of their environments. The relocation of products has to be scheduled while adhering to different time constraints. This paper addresses the challenge of developing solution approaches for such dynamic stacking problems in uncertain environments, specifically the environment represented by the "Hotstorage" simulation of the DynStack Competition 2023 that is part of the Genetic and Evolutionary Computation Conference. Building on existing approaches, two new solution approached are introduced: a score-based and a genetic solver. Comparing their performance with each other and with existing approaches shows that the score-based solver outperforms other solvers in regard to specific evaluation criteria while the genetic solver in its current form does not perform as well.

In agricultural water practices, accurate prediction of soil moisture (SM) is crucial for the efficient use and management of water resources. However, due to the complex structural characteristics and meteorological factors involved, establishing an ideal model for predicting SM can be challenging. Ensemble Methods (EM) offer a promising solution to overcome this challenge and increase the accuracy of SM prediction. In this study, we make a twofold contribution to the SM study regarding how various factors can impact the moisture content of the soil. First, we discuss the potential of the Extra Tree Regressor (ETR), CatBoost Regressor (CBR), Random Forest Regressor (RFR), Extreme Gradient Boosting Regressor (XGBR), and several other algorithms for SM prediction using atmospheric variables (i.e., air temperature, humidity, wind, solar radiance, rainfall, etc.) and soil factors (i.e., soil humidity, crop coefficient culture, evapotranspiration measured, evapotranspiration reference). We then look at the impact of contributing factors through sensitivity analysis. The findings indicate that the Model Stacking approach exhibited relatively superior performance (R2 = 0.9982) in predicting input variables, surpassing both Model Blending (R2 = 0.9969) and Model Ensembling of the three different machine learning models: ETR (R2 = 0.9966), RFR (R2 = 0.9929), and CBR (R2 = 0.9935). In general, the utilization of the Model Stacking technique exhibits encouraging outcomes in enhancing the accuracy of sentiment analysis predictions. This approach stems from its capability to address the constraints associated with rudimentary algorithms, characterized by restricted parameters and expedited computation durations. Moreover, the comprehensive analysis of feature importance underscores the critical role of humidity, wind dynamics, and other environmental variables in shaping soil moisture dynamics across diverse irrigation regimes.

Ant colony optimization (ACO) algorithms often prove effective and widely applicable for various problems, even in grid-based environments. However, from an outside perspective, the field of ACO and other swarm intelligence approaches may appear needlessly complex.This work should make the topic more approachable. It aims to enable newcomers and experienced researchers to quickly grasp and further develop solutions for ACO in grid-based environments. An open-source, Java-based simulation framework is presented, that allows other research or leisure-time projects to be implemented and executed more efficiently. Visualizing the resulting simulations can aid with understanding the basic principles and tweaking the underlying algorithms or parameters. Experiments have proven that the solution could produce pathfinding solutions in a grid environment using an example ACO-based algorithm. The corresponding algorithms are easy to implement using an abstract Java class that can be extended and used as a template. Furthermore, the visualization of the simulation is realized based on general principles that should make it intuitive and easy to understand. Future work based on this work should have all the necessary tools to create extensions or modifications as needed.

Teaching a database management course to industrial engineering students can be a difficult task, since the students are usually not especially interested in this non-core topic. The students show their low motivation by acting as very passive learners. Another complicating matter is that the course is taught online, which also increases the risk for dropouts. Therefore, we decided to address these two problems by introducing active learning elements in the course. In this paper, we present the current state of the database management course together with the active learning strategy incorporated into the teaching framework. This work has still to be considered as preliminary results, where we elucidate our first experimental findings.

Teaching a new object oriented programming (OOP) language nowadays is a challenging task. Many different approaches were proposed in the last years. But which one helps to achieve the optimal learning results? In the Nicomachean Ethics Aristotle wrote: “For the things we have to learn before we can do them, we learn by doing them.”. He already knew that the best way to learn something is to do it. Experience has confirmed the effectiveness of this approach, including in the field of programming. When learning a new programming language, it is extremely important to apply it directly to a specific task. Such a task could be, for example, the development of a well-known children's board game such as Nanu, Zicke Zacke etc. This is exactly the task we set our students as part of the “Object Oriented Programming with Java” module. In this paper we want to report the experiences we have made with this approach and make suggestions for the future.

In engineering courses laboratory sessions, there are long-established vehicles to transmit basic knowledge as well as engineering concepts and strategies. Even though their share is already high in most of technical degree programs, it is expected to increase further. The reason is that change is coming along with digitization and Industry 4.0. In this regard, laboratory work enables the students to practically implement complicated theoretical contexts to automation-technology and get familiar with real machines and equipment. On the other hand, with budget constraints in public education, laboratories are often criticized for being too expensive and too inefficient. Therefore it should always be a target to choose an efficient and adequate setup. To understand the institution's own needs, a project was carried out at Frankfurt University of Applied Sciences to analyze the situation and define strategies focusing on “laboratories for manufacturing technology”. A questionnaire was designed to assess motivational effects of laboratory work, and measure the level of knowledge gained. With respect to motivation, a clear preference could be found for laboratory sessions among the 76 participants. The knowledge gain showed some improvement from lab teaching, but all in all the effect was not as high as expected. This gave rise to further considerations and interesting hints on how to improve the operational setup.

If an engineer obtains correct conclusions, but cannot communicate these results in a proper way, then the value of the achievements could be diminishing dramatically. Awareness about the importance of acquiring soft skills in addition to technical knowledge is often present in engineering education, but the issue is still how the improvement of these key competencies should be addressed. Often our engineering classes are heterogeneous, both with respect to pre-knowledge of the mathematical and technical skills, but also concerning the cultural background of the students. This can be an additional challenge for the teaching situation. Instead of considering the cultural heterogeneity as an obstacle, we decided to use it as an advantage for enhancing soft skills in a mathematics course. In this paper, we describe a multicultural-based teaching approach for improving soft skills in engineering and computer science education.

An online B. Sc. program in mechanical engineering offers the students a flexible way to obtain a university degree in engineering, parallel to work and family life. However, the lecturers of the online program face the challenge to keep the students in the courses. The rate of dropouts can be reduced if the students participate actively in the course and feel responsibility and community with other students. These factors can be promoted by cooperative learning techniques, e.g. problem-based learning. This paper presents how problem-based learning was applied to an introductory laboratory class, taught completely online, for mechanical engineering students. We discuss the teaching framework of the laboratory class, as well as the results from the first in-class usage of the approach. Furthermore, a comparison with a previously used technique is given.

Video-sharing websites like YouTube are very popular, their usage can particularly be seen by younger people, the so called "Digital Natives". For example, about a billion of people visits YouTube every month and about a billion of hours is spent daily to watch videos there. In principle this enthusiasm for videos offers the opportunity to use them as a didactical tool. But even though the medium is available since decades, the usage of videos in didactics is still an expandable field and scientific publications are drawing a very complex scheme on this topic. Quantitative findings of the effect in teaching are very rare. Therefore, a team at the Frankfurt University of Applied Sciences carried out a project to analyze the situation and define strategies for the own needs focusing on the engineering subject of study "manufacturing technology". Based on literature and own considerations, a questionnaire was designed asking for preferences in learning media comparing chalkboard, power point and video based lectures. In parallel around 900 written examinations were analyzed over a period of 7 years. The analysis compared examination results on topics taught by video clips to results on topics taught in power point and chalkboard lectures. The concrete results did not show a significant advantage by video teaching, but gave interesting hints for follow-up projects. Regarding the different didactic tools, a clear preference could be found for video based lectures. Notable was that female students tended slightly to the chalkboard lectures.

It is a common held perception that nowadays young engineers and computer scientists suffer from a lack of mathematical knowledge, especially capabilities to cope with practical tasks. This becomes quiet obvious when they are confronted with the assignment to interact with each other while presenting their achievements. Nowadays, the math teacher faces complex issues like regarding the heterogeneity of the students' groups, the society's views towards mathematics, the overall workload, expectance, constant attendance or work behavior. In our Computer Science Study Program, the Calculus module was inserted starting with the winter term 2012/2013, additional to Algebra module for the first semester. The main reason was the missing knowledge in math basics towards the concepts needed for teaching other modules like Algorithms Analysis, Statistics, Optimization Techniques. Calculus was provided six times in our Computer Science Program and the results were increasingly better, since the content, methods and framework became a standard. Especially the students' evaluation in the first runs provided useful insights how to improve the teaching attempt. We present in this paper methods and tools, as well as contents and statistics to the topic, by visualizing the methods, tools and activities implemented in this large class approach.

Teaching a statistics course for undergraduate computer science students can be extremely challenging: As statistics teachers we are usually faced with problems ranging from a complete disinterest in the subject to lack of basic knowledge in mathematics and anxiety for failing the exam, since statistics has the reputation of having high failure rates. In our case, we additionally struggle with difficulties in the timing of the lectures as well as often occurring absence of the students due to spare-time jobs or a long traveling time to the university. This paper reveals how these issues can be addressed by the introduction of a blended learning module in statistics. In the following, we describe an e-learning development process used to implement time- and location-independent learning in statistics. The study focuses on a six-step-approach for developing the blended learning module. In addition, the teaching framework for the blended module is presented, including suggestions for increasing the interest in learning the course. Furthermore, the first experimental in-class-usage has been completed and the outcome is discussed.

Very often, there are complaints about low percentages of women in mechanical engineering studies. We may ask ourselves how this low percentage could be increased. There are lots of specific and scientific publications in the market drawing a very complex scheme on what to do. To complete this picture for our own measures, a team of the Department of Computer Science and Engineering at the FRA-UAS (Frankfurt University of Applied Sciences) organized a project with the focus on female students in our study programs. The survey results-apart from supposed points — were interesting and gave concrete hints how the teaching process could be improved. The survey stated, among other results, the big role of society's task to provide a wide range of information to girls for conducting an early socialization with technology and to generate more female role models. This cannot be plainly achieved by just one single institution. Another interesting observation not being covered by literature was the great impact of cultural aspects on female participation in technical aspects. For instance, the observations on the current teaching process to the republics of the former Soviet Union have shown that the image of “A woman's role in technology” is perceived completely different and as natural in these countries.

It is a commonly held perception that in general, engineers frequently suffer from a lack of soft skills. This becomes already quite obvious when prospective engineers are confronted with the task to communicate with their peers or while presenting their achievements in class. Insights from humanistic sciences are valuable to overcome such deficits at university level. Analysing specific texts, talking and exchanging opinions on the content with peers from inside and outside their subject area could enhance engineers' ability to efficiently communicate, for example. Fairy tales, as creative and cultural relevant texts with deep symbolic values, provide a rich ground to gain these capabilities. Specific skills, which could be improved by discussing fairy tales are, for example, realizing and decoding hidden meanings, associations and symbols. Also, creative thinking and the awareness of the importance of social and historical influences on content and understanding can be developed. Furthermore, differences in-between academic disciplines can be experienced by working in interdisciplinary, and often intercultural, teams. This approach is still in the process of field testing at Frankfurt University of Applied Sciences, now in the second run, where the interdisciplinary module Studium Generale offers space for soft skill training with a difference. So far, the results of this unusual approach to teach key competencies are promising for further projects. Nevertheless, an obstacle, this module still has to overcome, is the students', and especially prospective engineers' unwillingness to participate in such an “out-of-the-box”-setting without any reservations.

This paper describes an experiment for an interdisciplinary study at the Frankfurt University of Applied Sciences (FRA-UAS). It is a cooperative teaching approach of a professor for computer science and mathematics and one of communication sciences. All students of the university have to attend a module from the interdisciplinary studies pool and this means, the working groups are very heterogeneous with students from all the four departments: Architecture/Civil Engineering/Geomatics, Computer Science and Engineering, Business and Law, Social Work & Health. The link between these disciplines is that throughout all working environments - besides the required professional expertise - proper communication skills are mandatory. However, these soft skills include not only, for example, single extracts of knowledge about rhetoric or presentation technique training. Therefore, it is much more important to start with a rather holistic approach towards understanding the basic patterns of communication, in order to succeed in professional practice. Fairy-tales, as kind of creative and cultural relevant texts with deep symbolic value offer a rich factory to achieve several skills in the multidisciplinary education. Specific skills to improve are, for example, understanding and finding of denotative and connotative meanings, associations and symbols, innovative thinking, historical social approach, similarities to modern times, free retelling or discussing a given text, and team work.

The frequent ups and downs are characteristic to the stock market. The conventional standard models that assume that investors act rationally have not been able to capture the irregularities in the stock market patterns for years. As a result, behavioural finance is embraced to attempt to correct these model shortcomings by adding some factors to capture sentimental contagion which may be at play in determining the stock market. This paper assesses the predictive influence of sentiment on the stock market returns by using a non-parametric nonlinear approach that corrects specific limitations encountered in previous related work. In addition, the paper proposes a new approach to developing stock market volatility predictive models by incorporating a hybrid GARCH and artificial neural network framework, and proves the advantage of this framework over a GARCH only based framework. Our results reveal also that past volatility and positive sentiment appear to have strong predictive power over future volatility.

Natural swarms can have multiple structures and follow many different patterns or laws. Most recent studies try to obtain the best solutions in very specific and sophisticated contexts. In this research, instead, an application to find a reasonable approximation to the behaviour of any possible incoming environment is developed. This paper describes a basic framework, GUI and algorithms established in a way that can be easily modified and adapted to desired paradigms, parameters and rules. In experimental research, this will be able to provide help for many applications where natural swarm patterns are followed but no deep, expert simulator has yet been developed. To deal with the complexity of the biggest applications, the simulations are to be run in parallel computing using the MapReduce framework. Background and Motivation We were asked to develop a software which would fit the requirements for taking part in the InformatiCUP [1]. In this year's edition the participants were asked to find the best algorithms some robots should follow under a specific scenario [2]. These are deployed on the ocean's surface and can only communicate between other nearby boids, knowing not more than their relative positions. The aim is to move around the ocean's surface collecting manganese and gather together after a certain time or command. Many spreading or path algorithms could be appropriate, but there was a lack of information about what to optimize (time, fuel, unique track covered ...) so the goals were too ambiguous. None of the existing swarm algorithms that we found satisfied our expectations. As seen in previous studies [3][4][5], it's common to focus research in carefully delimited conditions, but no adaptable platform to use with our constrictions was found. The goal of our project switched to cover this lack of resources and try to set the basis for a lot of possible future developments and applications in experimental research. Different states and phases were introduced to modify the different algorithms' weights according to the desired behaviour of the robots at every moment. Section 2 thoroughly explains the context of the project and the first steps taken, describing how starting with different deployment patterns causes a need of very different movement rules, so regular shapes -square, circleand irregular deployments -random, Gaussian, combinedwere created to study all the possibilities. In section 3, we show how the application itself works, and how the iteration over the desired parameters can provide some acceptable results and how should they be understood for future applications. Section 4 introduces the further need of using parallel computing for the execution of the application. In section 5 the results are showed and commented to extract some conclusions leading to the future work purposed in section 6. Requirements and Basis Settled The requirements for the InformatiCUP were constrictive, especially when establishing which communications could take place between the robots. The uncertainty caused by the incomprehensibly ambiguous formulation of the task gave us total freedom to focus on the creation of a new versatile framework and leave constrictions behind. The required small viewing range and limited knowledge lead to the idea of applying swarm particle algorithms. Any other paradigm followed would be by intuition, and the time was limited to a few weeks which could not be spent on trying unfunded ideas. To keep the difference between what the robots could know and what not, a Simulator class was created to keep track and manage most of the information and share it properly with the other classes. The robot instances are stored in a HashMap also containing information about position (classes Robot and Coordinates). Simultaneously, a first visualisation of a of 500 by 500 positions was ALIFE 14: Proceedings of the Fourteenth International Conference on the Synthesis and Simulation of Living Systems implemented, so it was time to start with the deployment configurations. Optimising the gathering time is different to optimising the unique surface travelled or the fuel used after they start. Moreover, if the initial set of robots is positioned in different way, it becomes a key fact when it matters of which algorithm to follow in order to run over the surface following some criteria.. Hence, many deployment methods are available: DeployRandom(). Deploys an amount of robots in random positions within the given sea limits. Initially this algorithm was implemented to verify and debug the need of creating a new robot only within the viewing range from at least one other robot. Later in the project, this distribution stayed in the simulator for further testing of new features (new visualisation, determine movement behaviours) and as one of the starting point algorithms for the InformatiCUP. As it is a very inefficient starting set the challenge was more ambitious. DeploySquare() and DeployCircle(). Both deploy an amount of robots following uniformly filled squared or circle patterns. The first robot is one of the four central points of the figure, which grows in a spiral or surrounding shape from the centre point. DeployGauss() and DeployBadCenters(). Given the initial random centre (mean) and variance, deploy the robots following a Gaussian distribution, or in two unequal Gaussian distributions. This offers an interesting and challenging beginning, where two unequal groups are deployed close enough to see each other but far enough to, following the main algorithms seen later, tend to split up into two groups of boids. This is very useful when extreme conditions are going to be tested. Fig. 1. In order, Random, Square, Circle, Gauss and BadCenters deployments shown in the GUI.