We present a continuous integration and deployment (CI/CD) framework for Soccer Simulation 2D. On the one hand, we aim to share that system publicly with the community and, therefore, describe its components and their composition. On the other hand, we relate statistics produced by that system to concepts from reliability theory, which facilitates the derivation of reliabilty metrics for assessing a team’s defensive performance. In line with that, we present the results of a comprehensive case study in which the CI framework was utilized to find critical game situations, uncover weaknesses in the playing behavior of a soccer simulation team, and implement and evaluate appropriate counter measures.

We propose an architecture of a proof of concept for automated video summarization and evaluate its performance, addressing the challenges posed by the increasing prevalence of video content. The research focuses on creating a multimodal approach that integrates audio and visual analysis techniques to generate comprehensive video descriptions. Evaluation of the system across various video genres revealed that while video-based large language models show improvements over image-only models, they still struggle to capture nuanced visual narratives, resulting in generalized output for videos without a strong speech-based narrative. The multimodal approach demonstrated the ability to generate useful short summaries for most video types, but especially in speech-heavy videos offers minimal advantages over speech-only processing. The generation of textual alternatives and descriptive transcripts showed promise. While primarily stable for speech-heavy videos, future investigation into refinement techniques and potential advancements in video-based large language models holds promise for improved performance in the future.

The paper addresses the challenges and implications of (lacking) synchronization between agents in real-time multi-agent simulation systems. Based on two specific manifestations of mis-synchronization in 2D Soccer Simulation, termed holes and clashes, the paper makes two contributions: First, it provides a thorough historical review of the problem’s existence and urgency of occurrence, covering the time from the very early beginnings of the RoboCup initiative in the late 1990s till today. Second, it presents the results of an exhaustive empirical study that reveals quantitatively what impact synchronization problems may have on the playing performance of a simulated soccer team.

Video conferencing has become an invaluable tool to society and is essential in many professional and private settings. As the interaction of participants is purely virtual it lacks many aspects of “real life” including lack of or different feedback channels and henceforth the communication is less intimate. In particular, user engagement in video conferences is often difficult to assess for others and contributes to impediment of communication in general. This applies not only to small video conferences but even more so to conferences with a one to many topology. Measuring the user engagement in an accurate manner would provide a useful feedback channel to the speaker or organizer of a video conference and also enable many valued-added services such as predicting turn-taking for natural conversational speech intention and assisting AI-enabled video conference management. One of the challenges in measuring user engagement is to measure it outside an artificial lab setting and for an individual rather than an aggregate. In this work, we assess whether the problem of measuring user engagement is even well-defined. We also perform experiments with commercially available services and machine learning algorithms proposed by academia to test their ability of measuring user engagement in a realistic non-lab setting. Furthermore, we propose an outlier based algorithm and validate its superior performance against existing solutions.

The goal of the presented approach is to improve the stability of our RoboCup team code by providing an improved continuous integration software engineering process. As big and even small changes in our code base cannot be judged by just a couple of games, roughly 1000 games were run each night to have a good feeling whether changes were for the better or for worse. In addition, it is now possible to analyze the output to gain even deeper understanding of different approaches and parameters. This is supported by interactive visualization techniques. As a brute force approach will collect way too much data every night, it is necessary to condense the output and keep just a very small fraction of the detailed log data for further analysis. To decide which log files to keep different outlier detection algorithms are compared and optimized.

In this paper we introduce “SmartMail”, an add- on for Mozilla's e-mail client Thunderbird. It helps to increase productivity of users, especially the ones with international and multi-lingual e-mail recipients. We conducted a usability study which shows a significant acceleration of time spent writing e-mails as well as a reduced error rate for the chosen language and gender. After filling a questionnaire based on the System Usability Scale, the majority of test users rate the user experience satisfactory. Half of them even prefer using Thunderbird with this add-on over their own e-mail client.

Inter-agent communication has been playing an important role in soccer simulation 2D since its introduction. Its primary usage has been to communicate with teammates in order to share state observations to fill gaps in the players’ world models, to announce near future actions like passes or requesting passes, as well as for sharing and synchronizing on locker room agreements. In this paper, by contrast, our focus is on the communication of the opponent team. We present an approach for wiretapping and decoding opponent communication and systematically evaluate its impact. Our main finding is that a team that wiretaps its opponent and exploits intercepted information appropriately, can boost its own playing performance significantly.

In machine learning and numerical optimization, there has been an ongoing debate about properties of local optima and the impact of these properties on generalization. In this paper, we make a first attempt to address this question for case-based reasoning systems, more specifically for instance-based learning as it takes place in the retain phase. In so doing, we cast case learning as an optimization problem, develop a notion of local optima, propose a measure for the flatness or sharpness of these optima and empirically evaluate the relation between sharp minima and the generalization performance of the corresponding learned case base.

This paper summarizes the ongoing process of designing a wearable system that augments the learning environment by modeling learning content as competences, by providing feedback on how the user is acquiring these competences, and by enhancing the environment by Gamification elements. We evaluate the impact of the Interactive Learning Environment (ILE) in a preliminary user study. The application of design patterns facilitates the integration of any learning content, question types, or learning strategies. Extensions can be done at any time, without affecting the functionality of the existing application. An agile approach to platform selection (from Ionic to React Native) guarantees good technical results despite many uncertainties and difficulties encountered.

A remarkable feature of RoboCup’s soccer simulation leagues is their ability to quantify and prove the exact progress made over years. In this paper, we present and discuss the results of an extensive empirical study of the progress and the currently reached state of 2D soccer simulation. Our main finding is that the current decade has witnessed a continuous and statistically significant improvement of the overall level of play, but that the magnitude of the progress made has dropped clearly when compared to the previous decade. In accordance to this, we envision possible future prospects for the 2D league that might respond to our empirical findings.

We present a method for learning to interpret and understand foreign agent communication. Our approach is based on casting the contents of intercepted opponent agent communication to a bit-level representation and on training and employing deep convolutional neural networks for decoding the meaning of received messages. We empirically evaluate our method on real-world data acquired from the multi-agent domain of robotic soccer simulation, demonstrating the effectiveness and robustness of the learned decoding models.

The main focus of FRA-UNIted’s effort in the RoboCup soccer simulation 2D domain is to develop and to apply machine learning techniques in complex domains. In particular, we are interested in applying reinforcement learning methods, where the training signal is only given in terms of success or failure. In this paper, we describe the implementation of our newest behavior: Wiretapping and decoding opponent communication using convolutional neural networks in TensorFlow.

The automatic acquisition of a similarity measure for a CBR system is appealing as it frees the system designer from the tedious task of defining it manually. However, acquiring similarity measures with some machine learning approach typically results in some black box representation of similarity whose magic-like combination of high precision and low explainability may decrease a human user’s trust in the system. In this paper, we target this problem by suggesting a method to induce a human-readable and easily understandable – and thus potentially trustworthy – representation of similarity from a previously learned black box-like representation of similarity measures. Our experimental evaluations support the claim that, given some highly precise learned similarity measure, we can induce a less powerful, but human-understandable representation of it while its corresponding level of accuracy is only marginally impaired.

This paper presents a novel way to provide easy lecture recording. The actual recording is done on the hardware usually available anyhow: a laptop for the presentation and a smartphone. Voice and slides are recorded separately and displayed together in an HMTL5-based web application. This leads to minimal bandwidth requirements while still offering top notch quality of sound and graphics. Moreover the HTML5-based front-end can provide as a basis for collaborative learning and teaching.

This paper focuses on an investigation of case-based oppo- nent player modeling in the domain of simulated robotic soccer. While in previous and related work it has frequently been claimed that the predic- tion of low-level actions of an opponent agent in this application domain is infeasible, we show that { at least in certain settings { an online pre- diction of the opponent's actions can be made with high accuracy. We also stress why the ability to know the opponent's next low-level move can be of enormous utility to one's own playing strategy.

In this work we compare several new approaches for optimizing the emergency supply after a major incident online. For a given set of physicians, hospitals and transport vehicles the algorithms introduced in this work compute an assignment of arriving groups of casualties that suffer from specific types of injuries to available transport and medical capacities. We also consider how a patient's individual waiting time until medication impacts the corresponding course of disease and use the concept of penalty functions that can model casualty's health state. We use Simulated Annealing with transition probabilities favoring a balanced workload of vehicles and doctors. We show that using this optimization strategy in combination with a greedy initialization leads to better results compared to using only Greedy or D'Hondt assignment strategy which is currently used in practice.

Interaction data externalizes activities between an actor and an object, e.g. between a user and a computer system. The data is an important information source for user support and security applications. Nevertheless, generally little attention is given to interaction data. This paper emphasizes the role of interaction data usage for applications in the domain of human computer interaction and security. A consequent requirement is the clear and concise description of interaction data that enables the comparison of different systems that use interaction data. A scheme for classifying and describing processes related to the handling of interaction data is proposed. The key contribution of this paper is a classification scheme for interaction data and the description of three core processes required for interaction data management, namely interaction data collection, interaction data processing and interaction data organization. The scheme and the three processes are used to describe a concrete system for interaction data management at the computer desktop.