When constructing artificially intelligent systems, it may happen that bias and unfairness against a protected group, that do exist in the real world or in real-world data, are mapped into the digital world. If such systems are then used in practice for decisions that influence humans’ life, the obvious question is: Do the algorithms just pass through, perhaps mitigate, or even exacerbate such undesirable characteristics? That question has recently been addressed for the retrieve and retain phase of Case-Based Reasoning (CBR) systems. The paper at hand aims at providing an answer to that question for the reuse step in CBR, viz for case-based adaptation in k-nearest neighbor classification tasks.

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.

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.

The time required to retrieve a query's nearest neighbor may quickly become a CBR system's bottleneck when its case base contains a large volume of cases. Approximate retrieval techniques that build and employ complex index data structures can mitigate this issue by providing an acceptable tradeoff between time complexity and retrieval accuracy. In this paper, we propose a specialized multigraph-based index structure composed of multiple nested sub-case bases with individually labeled edge sets. We develop tailored algorithms for both constructing the index and utilizing it in a more efficient retrieval process, and we also evaluate our approach empirically using established benchmark datasets.

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.

The ability to correctly anticipate an opponent’s next action in real-time adversarial environments depends on both, the amount of collected observations of that agent’s behavior as well as on the capability to incorporate new knowledge into the opponent model easily. We present a novel approach to instance-based action prediction that utilizes graph-based structures for the efficiency of retrieval, that scales logarithmically with the amount of training data, and that can be used in an online and anytime manner. We apply this algorithm to the use case of predicting a dribbling agent’s next action in Soccer Simulation 2D.

Instance-based and case-based learning algorithms learn by remembering instances. When scaling such approaches to datasets of sizes that are typically faced in today’s data-rich and data-driven decade, basic approaches to case retrieval and case learning quickly come to their limits. In this paper, we introduce a novel scalable algorithm for both, the retrieval and the retain phase of the CBR cycle. Our approach builds an efficient graph-based data structure when learning new cases which it exploits in a stochastic any-time manner during retrieval. We investigate its characteristics both, theoretically and empirically using established benchmark datasets as well as a specific larger-scale dataset.

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.

The parameters of support vector machines (SVMs) such as kernel parameters and the penalty parameter have a great influence on the accuracy and complexity of the classification models. In the past, different evolutionary optimization algorithms were employed for optimizing SVMs; in this paper, we propose a social ski-driver (SSD) optimization algorithm which is inspired from different evolutionary optimization algorithms for optimizing the parameters of SVMs, with the aim of improving the classification performance. To cope with the problem of imbalanced data which is one of the challenging problems for building robust classification models, the proposed algorithm (SSD-SVM) was enhanced to deal with imbalanced data. In this study, eight standard imbalanced datasets were used for testing our proposed algorithm. For verification, the results of the SSD-SVM algorithm are compared with grid search, which is a conventional method of searching parameter values, and particle swarm optimization (PSO). The experimental results show that the SSD-SVM algorithm is capable of finding near-optimal values of SVMs parameters. The results also demonstrated high classification performance compared to the PSO algorithm.

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.

Path planning algorithms have been used in different applications with the aim of finding a suitable collision-free path which satisfies some certain criteria such as the shortest path length and smoothness; thus, defining a suitable curve to describe path is essential. The main goal of these algorithms is to find the shortest and smooth path between the starting and target points. This paper makes use of a Bézier curve-based model for path planning. The control points of the Bézier curve significantly influence the length and smoothness of the path. In this paper, a novel Chaotic Particle Swarm Optimization (CPSO) algorithm has been proposed to optimize the control points of Bézier curve, and the proposed algorithm comes in two variants: CPSO-I and CPSO-II. Using the chosen control points, the optimum smooth path that minimizes the total distance between the starting and ending points is selected. To evaluate the CPSO algorithm, the results of the CPSO-I and CPSO-II algorithms are compared with the standard PSO algorithm. The experimental results proved that the proposed algorithm is capable of finding the optimal path. Moreover, the CPSO algorithm was tested against different numbers of control points and obstacles, and the CPSO algorithm achieved competitive results.

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.

Grasshopper Optimization Algorithm (GOA) was modified in this paper, to optimize multi-objective problems, and the modified version is called Multi-Objective Grasshopper Optimization Algorithm (MOGOA). An external archive is integrated with the GOA for saving the Pareto optimal solutions. The archive is then employed for defining the social behavior of the GOA in the multi-objective search space. To evaluate and verify the effectiveness of the MOGOA, a set of standard unconstrained and constrained test functions are used. Moreover, the proposed algorithm was compared with three well-known optimization algorithms: Multi-Objective Particle Swarm Optimization (MOPSO), Multi-Objective Ant Lion Optimizer (MOALO), and Non-dominated Sorting Genetic Algorithm version 2 (NSGA-II); and the obtained results show that the MOGOA algorithm is able to provide competitive results and outperform other algorithms.

Fish identification is crucial for the survival of our threatened fish species. In this paper, a novel and robust biometric-based approach was proposed to identify fish species. The proposed approach consists of three phases. In the first phase, different features were extracted from fish images. In this phase, Weber's Local Descriptor (WLD) and color moments were used to extract texture and color features, respectively. Due to the high dimensionality of WLD features, in the second phase, Linear Discriminant Analysis (LDA) was applied to reduce the number of features and to discriminate between different classes. In the third phase, the AdaBoost classifier was used to identify fish species. We have collected a dataset that consists of four classes/species. To validate the results of the AdaBoost classifier, a comparison between three well-known classifiers (Naive Bayesian, k-Nearest Neighbor, and Multilayer Perception) was performed. The experimental results proved that our approach achieved excellent results (approximately 96.4%). Moreover, our model has been tested against different real challenges such as image rotation and image translation, and the proposed model achieved promising results.

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.

Measuring toxicity is an important step in drug development, and there is a high demand to develop computational models that can predict the drug toxicity risks. In this study, we used a dataset that consists of 553 drug samples that biotransformed in liver. The toxic effects were calculated for the current data are mutagenic, tumorigenic, irritant, and reproductive effects. The proposed model has two phases, in the first phase; sampling algorithms were utilized to solve the problem of imbalanced dataset, in the second phase, the Support Vector Machines (SVM) classifier was used to classify an unknown drug sample into toxic or non-toxic. Moreover, in our model, Dragonfly Algorithm (DA) was used to optimize SVM parameters such as the penalty parameter and kernel parameters. The experimental results demonstrated that the proposed model obtained high sensitivity to all toxic effects, which indicates that it could be used for the prediction of drug toxicity in the early stage of drug development.

Self-adaptation and the ability to assimilate new knowledge are two fundamental characteristics of intelligent systems. In this paper we leverage methods from evolutionary optimization and from case-based reasoning to construct an agent that is able to evolve in such a way that it is able to successfully master the popular video game Angry Birds.

Support Vector Machine (SVM) parameters such as penalty and kernel parameters have a great influence on the complexity and accuracy of the classification model. In this paper, Dragonfly algorithm (DA) has been proposed to optimize the parameters of SVM; thus, the classification error can be decreased. To evaluate the proposed model (DA-SVM), the experiment adopted six standard datasets which are obtained from UCI machine learning data repository. For verification, the results of the DA-SVM algorithm are compared with two well-known optimization algorithms, namely, Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The experimental results demonstrated that the proposed model is capable to find the optimal values of the SVM parameters and avoids the local optima problem.

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.