O Ethereum é uma plataforma de criptomoedas que permite a execução de contratos inteligentes, programas autônomos que operam em uma rede descentralizada. As vulnerabilidades nesses contratos representam grandes riscos financeiros e de segurança nos ecossistemas blockchain, motivando a automatização do processo de detectá-las. Este trabalho estuda a detecção de vulnerabilidades em contratos inteligentes Ethereum usando embeddings derivados de bytecode. Embeddings são representações vetoriais geradas por modelos de linguagem, que capturam as características estruturais de texto. Essas representações foram usadas como entrada para os algoritmos de regressão logística, árvore de decisão e floresta aleatória, com o fim de detectar quais contratos possuem vulnerabilidades. Os resultados mostram que os embeddings contêm informações úteis para distinguir contratos vulneráveis de não vulneráveis. O estudo também constata que a alteração da distribuição original dos dados durante o treinamento afeta significativamente o desempenho, destacando a sensibilidade das abordagens baseadas em embeddings às estratégias de amostragem.

Collaboratively maintained knowledge graphs like Wikidata rely on property constraints to detect data inconsistencies. This paper systematically formalizes potential repairs for Wikidata constraint violations, presenting a comprehensive taxonomy of repair strategies encompassing both instance-level (A-box) and terminological-level (T-box) changes. T-box repairs, which alter constraint definitions or Wikidata's class hierarchy, can simultaneously address multiple violations and, to the best of our knowledge, have not been investigated in detail before. We observe repairs over time and evaluate how specific patterns within our taxonomy are applied in practice. Our analysis of historical data reveals insights into the prevalence of repair patterns in Wikidata's collaborative environment. The results indicate that T-box repairs are particularly relevant for certain constraint types and the overall consolidation of Wikidata, where modifying constraint definitions can reduce the number of recurring violations.

Search as Learning (SAL) explores how users engage with search systems to acquire knowledge and develop understanding. Despite advances in SAL, the lack of general-purpose tools hinders reproducibility and standardization in experimental studies. This paper presents a framework to support researchers in designing SAL experiments, encompassing task creation, data collection, and learning assessment. To evaluate the proposal, we conducted a usability study with 12 participants, which yielded a score of 83.07, indicating excellent usability. Feedback of the participants also provided suggestions for improvement, guiding future development. This work contributes to strengthening methodological practices and fostering reproducibility in SAL research.

Embora o uso de Ambientes Virtuais de Aprendizagem (AVAs) permita capturar informações diversas sobre estudantes e cursos, o uso desses dados para geração de indicadores pertinentes à gestão é um desafio ativo em Learning Analytics. Ainda, a falta de uma plataforma gratuita que agregue diferentes indicadores é um limitador para o uso extensivo de dados dos AVAs na gestão universitária. Este artigo apresenta o desenvolvimento de um painel de indicadores educacionais para o ambiente Moodle, com foco na análise da trajetória estudantil em seis dimensões. A proposta integra contribuições teóricas e funcionais identificadas na literatura e em soluções existentes, organizando os dados de forma clara e acessível por meio de discretizações.

Background. The rapid growth of online data has made retrieving relevant information a challenging task, prompting the rise of Knowledge Base Question Answering (KBQA) systems that handle complex, multi-hop queries. Purpose. This extended work refines our previous pipeline by introducing structured dummy templates, a Hereditary Tree-LSTM (HTL) for classification, and more comprehensive analyses of entity recognition, property extraction, and SPARQL assembly. Methods. We enhanced the LC-QUAD 2.1 dataset with standardized templates and evaluated a flexible pipeline that integrates DeepPavlov, Falcon, SpaCy, qualifiers constraints, and reverse lookups. Results. Our experiments reveal that multi-tool entity recognition outperforms single-tool methods, while property extraction benefits from extended property sets and refined ranking strategies. Overall SPARQL correctness reaches up to 70–80% in mid-complex queries but remains lower in domain-specific subsets. Conclusion. The proposed synergy of NLP tools and refined dummy templates increases coverage for complex KBQA, though further improvements in morphological handling and specialized embeddings may be needed to address challenging multi-hop or niche queries comprehensively.

aaaa The adaptive learning community seeks to provide solutions to customize and enhance students’ learning experiences when accessing web-based learning systems. The adaptation usually occurs from the use of learning materials and user information data, which turns the adaptation process highly dependent on the quality of the repositories. Then, the best adaptation a system may offer might still not satisfy the users’ needs. In this work, we propose an approach to assist teachers and stakeholders in understanding repositories’ characteristics and their gaps according to students’ needs. Our approach, first, selects the best sequence of learning materials for each student, which is a well-known problem called Adaptive Curriculum Sequencing. Then, based on the selected sequences, we use optimization approaches, such as GRASP and Simulated Annealing, to generate new learning materials possibilities that can improve ACS recommendations. This way, our new approach assists teachers in assembling their learning materials. We have evaluated our approach by comparing it to a traditional approach using a real dataset, and the results are promising. In fact, it is possible to design customized materials using a combination of GRASP and brute force algorithms on the characteristics of the learning materials.

Context: The integration of IoT in healthcare has enhanced the capability of Ambient Assisted Living by enabling reliable real-time health monitoring, particularly for heart failure identification. Predictive models play a crucial role in identifying potential heart failures, improving patient outcomes through continuous monitoring and analysis. Problem: Traditional predictive models rely on centralized servers, facing issues like network latency, disruptions, and data overload. These challenges hinder real-time health data processing, limiting their ability to provide timely heart failure risk predictions. Solution: This study proposes a framework that embeds machine learning models directly into mobile devices, leveraging edge computing for real-time heart failure risk assessment. By processing data locally, the solution reduces latency, enhances reliability, and ensures greater data privacy while maintaining predictive accuracy.IS Theory: The framework aligns with socio-technical IS theory by integrating technical innovations and user-centric needs. Embedding predictive models into mobile devices enhances real-time predictions and patient care, bridging technology and human interaction. Method: A heart failure prediction dataset was utilized, employing supervised classification algorithms—Random Forest, K-Nearest Neighbors (KNN), and Logistic Regression. Data preprocessing included handling missing values, feature scaling, and encoding. The trained models were deployed on mobile devices using ONNX Runtime for efficient real-time inference. Summary of Results: The proposed system successfully performed real-time heart failure risk prediction on mobile devices, achieving competitive accuracy. Random Forest outperformed other models, reaching an accuracy of 85.33%, demonstrating its effectiveness in edge computing environments. The approach significantly mitigates the connectivity and latency challenges of centralized systems while enhancing data security. Impact in the IS Area: This research highlights the potential of edge computing to enhance real-time healthcare applications by reducing reliance on cloud infrastructure. Future works includes conducting real-environment patient evaluations to validate the system’s clinical applicability, integrating biomedical sensors, improving predictive models using deep learning, and exploring federated learning for privacy-preserving model training.

This paper aims to clarify the characteristics of Digital Twins (DTs) in their most advanced conceptual development, Cognitive Digital Twins (CDTs), and analyze their support for the implementation of the Circular Economy (CE). A systematic literature review was conducted using a specially developed five-dimensional analytical framework to characterize DT proposals and their potential for CE based on an established framework for circularity strategies. The study indicates that cognitive and hybrid DT approaches tend to cover high levels of interoperability, data flow, system levels, and cognitive processes. However, CDT use in CE demands harmonizing different strategies to cover the complete product lifecycle, which recent research on DTs has not fully addressed. This study is the first to systematically review cognitive digital twins and their relation to circularity, offering an analytical framework that can be expanded for future research in various application areas of Industry 5.0.

This paper presents a Genetic Algorithm (GA) to automate the generation of reading fluency assessment items, reducing manual effort while meeting pedagogical constraints. Candidate solutions are sequences of words optimized by a multi-objective function that penalizes constraint violations and repetitions. Constraints include canonicity, syllabic variety, grapheme presence, and prosodic continuity. Experiments show that the GA effectively produces valid word lists, with larger populations yielding faster and more stable convergence. A 5% mutation rate was sufficient to preserve diversity. The method is flexible, scalable, and aligned with educational standards.

A educação online exige abordagens personalizadas que respeitem as particularidades de cada estudante. Assim, o Sequenciamento Curricular Adaptativo (ACS) surge como uma solução promissora, mas sua aplicação ainda é limitada. Este estudo analisa criticamente abordagens de ACS, explorando sua adaptação em ambientes virtuais de aprendizagem, com foco no Moodle. Realizou-se um levantamento dos principais dados e métodos utilizados em ACS, mapeando esses elementos com recursos nativos do Moodle. Como resultado, propõe-se uma prova de conceito que demonstra a viabilidade de percursos adaptativos na plataforma. A pesquisa contribui para transformar modelos teóricos em soluções práticas, avançando para um ensino personalizado.

Context. Harmful Code denotes code snippets that harm the software quality. Several characteristics can cause this, from characteristics of the source code to external issues. By example, one might associate Harmful Code with the introduction of bugs, architecture degradation, and code that is hard to comprehend. However, there is still a lack of knowledge on which code issues are considered harmful from the perspective of the software developers community. Goal. In this work, we investigate the social representations of Harmful Code among a community of software developers composed of Brazilian postgraduate students and professionals from the industry. Method. We conducted free association tasks with members from this community for characterizing what comes to their minds when they think about Harmful Code. Then, we compiled a set of associations that compose the social representations of Harmful Code. Results. We found that the investigated community strongly associates Harmful Code with a core set of undesirable characteristics of the source code, such as bugs and different types of smells. Based on these findings, we discuss each one of them to try to understand why those characteristics happen. Conclusion. Our study reveals the main characteristics of Harmful Code by a community of developers. Those characteristics can guide researchers on future works to better understand Harmful Code.

Context: Information seeking plays a key role in the learning process, enabling individuals to acquire knowledge and make well-informed decisions. However, this process is not exempt from cognitive biases that can distort the way we interpret and use available information. Ongoing research seeks to comprehend and mitigate these biases to enhance search efficacy and promote effective learning. Problem: Despite these efforts, existing empirical experimentation remain confined to isolated platforms, hindering reproducibility and collaborative progress within the field. This limitation underscores a critical need for a more unified approach to experimentation. Solution: In response, we propose a comprehensive framework designed to support and standardize experimentation. IS theory: Our approach aligns with Design Theory, establishing a connection between cognitive biases and the technical dimensions of the information system. Method: To define the requirements of the proposed framework, a thorough literature review on cognitive biases in search was conducted. The framework's efficacy is demonstrated through a proof of concept. Summary of Results: We showcase the framework applicability by instantiating it with a study on confirmation bias within a health-related search task. This implementation is particularly relevant as it integrates crucial components and requirements identified in previous research. Contributions and Impact in IS area: Our proposed framework bridges a significant gap in the field by presenting a standardized approach to conducting experiments on information seeking and cognitive biases. This not only fortifies the reliability of individual studies but also fosters collaborative efforts, enabling a more profound understanding of information-seeking behaviors across diverse domains within the Information Systems community.

Gamification promotes user engagement with software features through the incorporation of game elements and rules. Gamification is often incorporated a posteriori into already existing systems. Success in gamifying an existing system depends on careful planning and the evolution of its previously produced software artifacts. This is particularly true for artifacts produced in the earliest development phases, such as the requirements specification. Incorporating game elements and rules into an existing system is far from trivial. Developers eventually struggle with performing certain development activities, such as evolving existing requirements and selecting game elements and game rules. This paper reports our practical experience in gamifying an existing system. Based on this experience, we introduce Gamify4Fun, a method that aims to assist developers in performing some key activities to gamify existing systems. We built Gamify4Fun based on the experience of 15 developers involved with the gamification of an existing healthcare system. We started by adopting an original method aimed at gamifying systems being built from scratch. As we needed to adapt the original method for supporting the particularities of gamifying a previously developed system, we refined the original method’s development activities and their respective phases. We also interviewed the developers to capture their perception of challenging development activities through gamification. The interviews’ outcomes guided some further refinements to the original method. Gamify4Fun supports the gamification of existing systems at the earliest development phases: from the preparation of the system gamification to both the gamification and system design. We refined the development activities of the original method as much as needed; we also used or adapted the activities prescribed by other methods from the literature, in the context of gamification from scratch, to fill gaps whenever necessary. By reporting our practical experience and introducing a gamification method, we expect to guide development teams in gamifying their existing systems, as well as shed insights about the current, unaddressed limitations of existing approaches (including ours) to gamifying existing systems.

Exception handling is a technique used to manage errors or exceptional events that occur during program execution and disrupt the standard flow of the program. Although this method is beneficial, developers often neglect proper exception handling. Misuse of exception handling mechanisms can lead to exception handling anti-patterns in the source code, negatively impacting software quality. Additionally, many projects nowadays are built using multiple languages, which increases the complexity of identifying these anti-patterns and the workload involved in configuring various linters or detection tools. To address this gap, we developed a tool that identifies exception handling anti-patterns in three popular languages: Java, TypeScript, and Python. To evaluate our tool, we conducted an empirical study on 11 multi-language projects to investigate the occurrence of exception handling patterns across different languages. Our results demonstrated that the occurrence of exception handling anti-patterns is similar between Python and TypeScript. However, in Java, exception handling anti-patterns occur up to four times more frequently in the analyzed projects. The tool is available on GitHub1, along with a video2 demonstrating its use.

Robustness is critical for ensuring that software functions correctly under adverse conditions. Exception-handling mechanisms in programming languages enable developers to deal with these adverse conditions. However, implementing exception-related code can present significant challenges to developers. We investigated exception-related code contributions across Java projects in the Apache ecosystem. We analyzed exception-related pull requests (exception-PRs), which were detected using a validated heuristic. We produced a comprehensive dataset of 988 exception-PRs. We observed no statistically significant differences in complexity metrics between exception-PRs and non-exception-PRs. We also found no significant differences in developers' behavior metrics, indicating consistent engagement regardless of whether the pull request addressed exception-related code. A manual analysis revealed that most exception-PRs focused on system improvements rather than bug fixes, suggesting proactive efforts to enhance software robustness. Moreover, the most frequently addressed aspects of exceptional code in these exception-PRs were: (i) the external representation of adverse situations to end-users (more than 40% of the PRs) and (ii) the implementation of effective error-handling actions (nearly 35% of the PRs) to promote program recoverability. Interestingly, a significant proportion of exception-PRs simultaneously addressed multiple aspects. By understanding the nature and characteristics of exception-PRs, we expect to better support developers in managing erroneous conditions and improving software robustness.

The exponential growth of data on the internet has made information retrieval increasingly challenging. Knowledge-based Question-Answering (KBQA) framework offers an efficient solution that quickly provides accurate and relevant information. However, these frameworks face significant challenges, especially when dealing with complex queries involving multiple entities and properties. This paper studies KBQA frameworks, focusing on improving entity recognition, property extraction, and query generation using advanced Natural Language Processing (NLP) and Artificial Intelligence (AI) techniques. We implemented and evaluated combination tools for extracting entities and properties, with the combination of models achieving the best performance. Our evaluation metrics included entity and property retrieval, SPARQL query completeness, and accuracy. The results demonstrated the effectiveness of our approach, with high accuracy rates in identifying entities and properties.

In this paper, we delve into the crucial role of constraints in maintaining data integrity in knowledge graphs with a specific focus on Wikidata, one of the most extensive collaboratively maintained open data knowledge graphs on the Web. The World Wide Web Consortium (W3C) recommends the Shapes Constraint Language (SHACL) as the constraint language for validating Knowledge Graphs, which comes in two different levels of expressivity, SHACL-Core, as well as SHACL-SPARQL. Despite the availability of SHACL, Wikidata currently represents its property constraints through its own RDF data model, which relies on Wikidata’s specific reification mechanism based on authoritative namespaces, and – partially ambiguous – natural language definitions. In the present paper, we investigate whether and how the semantics of Wikidata property constraints, can be formalized using SHACL-Core, SHACL-SPARQL, as well as directly as SPARQL queries. While the expressivity of SHACL-Core turns out to be insufficient for expressing all Wikidata property constraint types, we present SPARQL queries to identify violations for all 32 current Wikidata constraint types. We compare the semantics of this unambiguous SPARQL formalization with Wikidata’s violation reporting system and discuss limitations in terms of evaluation via Wikidata’s public SPARQL query endpoint, due to its current scalability. Our study, on the one hand, sheds light on the unique characteristics of constraints defined by the Wikidata community, in order to improve the quality and accuracy of data in this collaborative knowledge graph. On the other hand, as a “byproduct”, our formalization extends existing benchmarks for both SHACL and SPARQL with a challenging, large-scale real-world use case.

Although the disappearance of individuals is not a recent phenomenon, it remains a prevalent issue that inflicts significant emotional distress upon the families of the missing. Unfortunately, state action about this matter is lacking in several countries. One promising approach to address this problem involves appealing for information and reaching out to a wider network of individuals who may possess the ability to assist in locating the missing person. Social media platforms, such as Twitter, have proven to be particularly effective in disseminating information. However, the effectiveness of information dissemination is crucial to raise awareness within the community as a whole. This paper presents a method for identifying influential individuals on Twitter, with a focus on their geographic location, to maximize the diffusion of information about missing persons. Given the significance of the social circles and communities associated with the disappeared individuals, incorporating location data becomes an essential feature in the missing person domain. The contribution of this paper is threefold: (i) a novel method to identify location-aware influencers on Twitter based on an operational research model, (ii) an analysis of the information dissemination using publicly available missing person data collected from Brazilian non-governmental organizations and state websites, and (iii) a new missing person dataset that can serve as a valuable resource for further research.

Objective: To identify important aspects for the analysis of digital twins and discuss digital twins created for the Brazilian industry based on characteristics of this technology. Methodology: A literature review of digital twins definitions and analysis of digital twins created for the Brazilian industry considering five analysis dimensions. Originality/relevance: Proposition and use of a set of dimensions to analyze digital twins, which can be used by researchers and practitioners for better understanding and characterizing digital twins proposals. Results: Considering the digital twins created for the Brazilian industry, we found these proposals distinguish in some aspects, mainly the data flow between physical and digital objects, system level, and cognitive capabilities, but aspects such as interoperability, cognitive processes, and life cycle are uncovered in these digital twins. These aspects, however, are responsible for the innovation and disruption digital twins can provide to the Industry. Contribution: Definition of dimensions to analyze digital twins and evidence of the presence and absence of some characteristics in digital twins created for the Brazilian industry.

Adaptive Curriculum Sequencing (ACS) is an important issue in personalized learning. In ACS problems, one desires the best sequence of learning materials that meet the profile of a given student. To do so, multiple features of the students and the materials used are necessary to generate good solutions. In fact, understanding the students’ goals, motivation, and preferences is not an easy task and, consequently, different Internet of Things (IoT) approaches to gather this information during the learning process have been proposed. Actually, some works from the literature consider five objectives and, in this case, one has a many-objective optimization problem. Instead of solving the optimization problem considering the multiple objectives individually, the usual approach is to obtain solutions for a weighted sum of the objective values using search approaches for mono-objective optimization problems. However, this kind of approach may bias the search and limits the capacity of finding good results. Here, we solve the multi-objective ACS problem considering five objective functions. NSGA-II, a well-known Genetic Algorithm for multi-objective optimization problems, was used. In addition, the aggregation trees were employed to reduce the number of objectives to two and three due to the large number of objectives in the original problem. ACS problems from the literature were used to comparatively evaluate the proposed methods and the results obtained were compared to those found by the traditional approach of summing the objective values. According to these results, the best curriculum sequences were reached when using the proposal.