This work presents a framework for risk-aware pathfinding in topographic terrains, designed to support realistic and tactical route planning in agent-based simulations. The proposed method enhances the A ∗ algorithm by incorporating topographic elevation and risk costs into its path cost estimation function, enabling the computation of paths that reflect tactical constraints in simulation scenarios. To improve the efficiency of heuristic estimations used in path search, we introduce the use of Deep Neural Networks (DNNs) trained to learn correction factors for heuristic functions. These models adjust distance and risk estimates during the search, outperforming conventional heuristic functions, particularly in complex environments. To reduce the computational overhead typically associated with DNN inferences, we use an Iterative Heuristic Map (IHM), a caching and batching mechanism that minimizes redundant model queries by reusing heuristic estimates for spatially related nodes. The framework introduces the notion of View Point Configurations (VPCs), a method to model multiple risk sources in a digital elevation model (DEM) by simulating the field of view from elevated risk points. VPCs enable the representation of dangerous and safe terrain areas and are used to train and evaluate DNN models. The presented work conducted comprehensive experiments in increasingly complex scenarios, including unitary, mixed, and multiple active VPC configurations. It also analyzes the risk exposure of the paths calculated by the proposed framework, comparing them with those generated by the baseline A ∗ algorithm. The results confirm the effectiveness of the framework in reducing computation costs, enhancing risk-aware DNN generalization, and improving tactical pathfinding in topographic terrains.

Deep learning (DL) electrocardiogram (ECG) predictions are highly valuable for advancing early diagnosis, risk assessment, and treatment planning for cardiovascular diseases. For developing effective predictive models, ECG synthetic data generation is a promising solution for enhancing training datasets while ensuring patient privacy. This work investigates the use of Transformer-based Generative Adversarial Networks (GANs), specifically the TTS-CGAN model, to generate realistic synthetic ECG time-series data. The TTS-CGAN model architecture relies on Transformer encoders for both the generator and discriminator, incorporating state-of-art self-attention mechanisms to enhance time-series data modeling. The research contributions include (i) the analysis of TTS-CGAN for ECG synthesis, (ii) the assessment of synthetic data quality using both MIMICIV and MIT-BIH datasets, (iii) the key introduction of hybrid datasets combining real and synthetic ECG signals to assess model training performance, and (iv) the evaluation of generated data using qualitative and quantitative metrics. Experiments demonstrate that synthetic ECG signals maintain statistical properties similar to real data, with high Cosine similarity (up to 0.99) and low divergence (Jensen-Shannon distance 0.2). Hybrid dataset experiments reveal that models trained on a mix of real and synthetic data (up to 60 % synthetic) retain classification performance while improving dataset diversity.

This paper investigates synthetic data generation through Generative Adversarial Networks (GANs) and Transfer Learning (TL), focusing on Chronic Kidney Diseases (CKD). It analyzes whether GANs, particularly the medGAN and CorGAN architectures, can generate high-quality synthetic tabular data and how TL can enhance this process. The contributions include evaluating alternative medGAN and CorGAN setups, incorporating WGAN and WGAN-GP loss functions, and assessing how fine-tuning with TL impacts data realism and classifier performance. The models were pre-trained on a larger CKD dataset and fine-tuned on a smaller one, using consistent hyperparameters with reduced learning rates during fine-tuning. Experiments involved training Random Forest classifiers in various settings: using real data, synthetic data, and a combination of both, with and without TL. Metrics like dimension-wise probability and multiple training/testing scenarios are employed to assess the quality and utility of the generated data. The synthetic data, especially when generated using TL, improved classifier performance significantly in scenarios where training and testing datasets differed. Notably, F1-scores improved by up to 74.3 % when using TL-generated data. These findings support the use of GANs with TL as a powerful approach to overcome data limitations in healthcare research.

Existing adversarial Domain Generation Algorithms (DGAs) often fail to produce realistic synthetic domains because their sequential models cannot adequately capture complex character dependencies within domain names. This paper presents TITAN DGA, a new approach that addresses this limitation by integrating a Generative Adversarial Network (GAN) with a Transformer-based autoencoder, enabling the model to effectively learn long-range dependencies to generate synthetic domains. The solution incorporates adversarial self-augmentation to enhance evasiveness through iterative retraining. The model learns from its own generated samples using two strategies: a broad approach that improves overall generation quality, and a targeted approach that specifically leverages previously successful evasions to refine the model’s adversarial capabilities. Comprehensive evaluation against multiple state-of-the-art classifiers and competing adversarial DGAs demonstrates that TITAN DGA achieves superior evasion rates while maintaining realism.

Network datasets are essential resources to evaluate the effectiveness of new technologies in diverse network environments. Unfortunately, these datasets are often not publicly available or lack the necessary density and diversity for thorough testing. Generative Adversarial Networks (GANs) have shown promise in generating realistic synthetic data. Transfer Learning (TL), which transfers knowledge from one domain to another, is another technique that potentially enhances data generation. This paper combines such techniques in a novel framework capable of generating synthetic network data tailored to specific desired protocol types. A GAN is pre-trained on an available dataset with substantial size and diversity corresponding to a source protocol. Then, exploring a TL method, the GAN undergoes fine-tuning using a smaller target protocol dataset. This fine-tuning allows the GAN to generate an augmented dataset containing relevant samples from the target domain. Developed experiments include two different data generation scenarios: i) Intra-protocol: transferring knowledge from datasets with different characteristics, considering the same source and target protocols; ii) Inter-protocol: transferring knowledge between protocols from different layers of the ISO-OSI reference model. The obtained results demonstrate that the proposed GAN and TL framework effectively generates high-quality synthetic traffic, featuring a significant rate of well-formed packets; a high percentage of packets with an appropriate query response; and a considerable similarity between the generated and real packets - measured by the FID (Fréchet Inception Distance) - compared to standard GANs without fine-tuning.

The dynamic and increasing sophistication of cyberattacks and vulnerability exploitation creates a need for Explainable Artificial Intelligence (XAI) approaches that help maintain cyber resilience in organizations. In structured cybersecurity incident management, effective incident response demands explainable outputs from AI-based decision-support systems. To approach this problem, this work presents a framework for reusing concrete experiences of cybersecurity incident response, capturing problem-solving data and knowledge as cases for integrated Case-Based Reasoning (CBR) and Clustering. The contribution includes cluster-based query answer analysis, where cybersecurity analysts reuse clusters of retrieved incident response cases to build answers to new problems. Clustering helps analysts identify relevant groups from ranked lists of retrieved cases, making the reuse process more structured and understandable, especially when dealing with retrieval results for broad and ambiguous queries. Different clustering methods are applied to organize retrieved incident response cases from a case base, supporting the grouping of similar cases for a more straightforward interpretation. Multiple experiments, including cross-validation and real-world incident response testing, are conducted to demonstrate the effectiveness of the proposed framework in improving the decision-support system’s precision. The results indicate that exploring cases and clusters can enhance the selection of incident response procedures for reuse, mainly when analysts identify the most relevant clusters of retrieved cases for the given problem situations. The proposed framework contributes to the organization and understanding of responses to cybersecurity incidents, besides supporting more informed decision-making, ultimately improving cybersecurity incident management.

Computer games continue to present challenging experimental fields for developing Artificial Intelligence (AI) models. With Case-Based Reasoning and Clustering, this work proposes novel cases and clusters-based reuse criteria for implementing card-playing agents with diversified playing skills. Using the game of Truco, a common game in South America, we detail how game actions are reused from past cases selected as query answers for given game problems. In doing so, the majority rule, the probability-based lottery, the probability of victory, and the number of points won reuse policies are used to select a cluster of cases. Then these policies are also used to select game actions from the cases within the selected cluster. Investigating the combined exploration of reuse policies, experiments of different natures evaluate the performance of implemented Truco bots disputing matches against each other. Players in these tournaments are equipped with varied policies and use case bases constructed differently.

Simulation systems are important tools for the development of training tasks in different application domains. The use of military simulated problems (MSPs) in simulation systems allows the introduction of challenging decision making situations to military trainees so that they can exercise their skills. To support the implementation of such MSPs in an artillery battery simulator, this work investigates how to compute dynamic changes, such as flooding-driven modifications, into the hierarchical and irregular navigation grid structure expressing the environmental characteristics of large virtual terrain scenarios. The paper describes an efficient mechanism to process these modifications during simulation runtime, without affecting the overall system response time, thus the user interaction with simulations. Experimental results show that the proposed techniques are able to handle time-constrained modifications of large virtual terrain scenarios.

This work investigates the combination of cases and clusters in the reuse of game actions (e.g., cards played, bets made) recorded in the cases retrieved for a given query in Case-based Reasoning (CBR) card-playing agents. With the support of the K-MEANS clustering algorithm, clustering results detailing problem states/situations and game outcomes relationships recorded in cases from the case base guide the execution of augmented reuse policies. These policies consider the game actions recorded in the retrieved cases in the selection of the clusters to be used. Then, the cases that belong to the selected clusters are used in the determination of which game action is reused as a solution to the current game problem situation. With this two-step reuse process, the proposed policies rely on the majority with clusters, the probability with clusters, the number of points won with clusters and the chance of victory with clusters. To evaluate these proposals, card-playing agents implemented with different reuse policies competed against each other in duplicated game matches where all of them played using the same set of cards.

Intelligent computing techniques have a paramount importance to the treatment of cybersecurity incidents.In such Artificial Intelligence (AI) context, while most of the algorithms explored in the cybersecurity domain aim to present solutions to intrusion detection problems, these algorithms seldom approach the correction procedures that are explored in the resolution of cybersecurity incident problems that already took place.In practice, knowledge regarding cybersecurity resolution data and procedures is being under-used in the development of intelligent cybersecurity systems, sometimes even lost and not used at all.In this context, this work proposes to integrate Case-Based Reasoning techniques and IODEF standard in order to retain concrete problem-solving experiences of cybersecurity incident resolution to be reused in the resolution of new incidents.Experimental results so far obtained with a Case-based Cybersecurity Incident Resolution System (CbCSecIRS) implemented show that information security knowledge can be retained in a reusable memory, so improving the resolution of new cybersecurity problems.

Intelligent computing techniques have a paramount importance to the treatment of cybersecurity incidents. In such Artificial Intelligence (AI) context, while most of the algorithms explored in the cybersecurity domain aim to present solutions to intrusion detection problems, these algorithms seldom approach the correction procedures that are explored in the resolution of cybersecurity incident problems that already took place. In practice, knowledge regarding cybersecurity resolution data and procedures is being under-used in the development of intelligent cybersecurity systems, sometimes even lost and not used at all. In this context, this work proposes the Case-based Cybersecurity Incident Resolution System (CCIRS), a system that implements an approach to integrate case-based reasoning (CBR) techniques and the IODEF standard in order to retain concrete problem-solving experiences of cybersecurity incident resolution to be reused in the resolution of new incidents. Different types of experimental results so far obtained with the CCIRS show that information security knowledge can be retained with our approach in a reusable memory improving the resolution of new cybersecurity problems.

We have put forward an approach to online collaborative discussion of software development problems based on Argumentation theory.Having records of past discussions can significantly help solve problems in new projects, and CBR techniques are used to retrieve the most similar cases.However, long discussions on past projects still contain too much information to provide support in new discussions.To address this problem, in this paper we introduce the idea of explanation templates that are able to summarize past experiences, particularly for risk management discussions.We formalize this notion of template, introduce the main templates we have developed to support explanation of past experience with risk management, and report the results of a case study on a realworld software project to assess the usefulness of those templates.

Agent behaviors in simulation systems are related to fundamental capabilities of realistically developing (semi) autonomous navigation actions. This is particularly important when dealing with the implementation of Computer Generated Forces (CGFs) for simulation systems in tactical military training applications. Moreover, these systems take into consideration the particularities of the domain-specific simulation tasks and the numerous heterogeneous CGFs inserted on them in order to generate better knowledge and learning experience to simulation system users. Based on these reasons, this paper reviews recurrent navigation problems as to propose a task-oriented and parameterized (semi) autonomous navigation framework to deal with CGF navigation needs in military simulation. Combining global and local navigation techniques, and controlled transition between alternative degrees of navigation autonomy, the framework aims to overcome the challenges of implementing customizable CGF navigation behaviors and, at the same time, to allow interaction with both users and other simulation systems in distributed simulation settings. A case study is presented in which the proposed techniques are analyzed in a domain-specific simulation problem providing evidence of their suitability to address the studied military simulation problems.

The engagement of project stakeholders in collaborative debates of risk management has an important contribution to software projects.To promote the identification, (re)use and critical analysis of stakeholders' arguments in these debates, this paper lays out a knowledge engineering process for the development of "argumentation schemes" for risk management.This process covers activities of identification, interpretation and causal-and-effect analysis of typical risk statements.From such risk management information and reusing generalized argumentation templates from argumentation catalogues discussed in the field of Artificial Intelligence, the process leads to the specification, generalization, validation and indexing of the developed schemes.As implemented in our project, a web-based system to support the execution of these development activities allows the recording of these schemes in a semi-structured representation format.An argumentation scheme for risks of non-stable requirements is presented so as to show the reusable argumentation artifacts that can be produced when our development process is followed.

Managing risks in real-world software projects is of paramount importance.A significant class of such risks is related to the engineering of requirements, commonly involving the presentation and analysis of risk management arguments from both software engineers and clients involved in collaborative debates.In this work, drawing inspiration from argumentation theory in Artificial Intelligence, we introduce a number of "argumentation schemes" and associated "critical questions" to support such discussions.In doing so, we propose schemes related to risks due to excessive numbers of requirements; inadequate client representatives and poor understanding of client needs; incorrect, incomplete and conflicting requirements, and complex and non-traceable requirements.We also present a case study where the developed schemes were used to support the discussion of requirement risks in the context of a research and prototyping software project for the Brazilian Army.

Managing risks in real-world software projects is of paramount importance. A significant class of such risks is related to the engineering of requirements, commonly involving the presentation and analysis of risk management arguments from both software engineers and clients involved in collaborative debates. In this work, drawing inspiration from argumentation theory in Artificial Intelligence, we introduce a number of “argumentation schemes” and associated “critical questions” to support such discussions. In doing so, we propose schemes related to risks due to excessive numbers of requirements; inadequate client representatives and poor understanding of client needs; incorrect, incomplete and conflicting requirements; complex and non-traceable requirements; non-stable requirements; and low quality requirements. We also discuss a case study and two experiments where the developed schemes supported the discussion of requirement risks in software projects. The overall results of these experiments indicate that our schemes are useful in the identification, proposition and analysis of requirement risks, adequately supporting debates on requirement risks.

Risk management aims to discuss the probabilities and consequences of risks on the goals of a software project.In such projects, there are dependence relationships between risks, although they are not treated yet by standard risk management practices.This paper is concerned with the analysis of risk dependencies, where these risks are assessed when multiple project stakeholders are involved in the development of collaborative risk debates.Our approach is based on a dialogue game protocol for collaborative risk management.This protocol mediates not only the discussion tasks of risk identification, risk analysis and risk planning, but also the collaborative debate regarding the identification and treatment of dependent risks.These risk management concepts are represented in a Bayesian network model for a risk management discussion situation, where alternative simulation scenarios can be proposed and tested in this probabilistic model according to discussion participants' requests.As observed in a case study, results from this process lead to the enhancement of the argumentative analysis of risk management issues developed by project stakeholders.