Context: Generative models are nowadays widely used to generate graphical content used for multiple purposes. However, it has been shown that the images generated by these models could reinforce societal biases already existing in specific contexts. The Software Engineering (SE) community is not immune to gender and ethnicity disparities, which could be amplified by the use of these models. Hence, if used without consciousness, artificially generated images could reinforce these biases in the SE domain. Objective: In this paper, we focus on understanding the implicit bias exposed by general-purpose open-source image generation models towards SE tasks. In addition, we investigate the extent to which it is possible to mitigate the bias by using prompt engineering techniques. Methods: We perform an extensive empirical evaluation of the implicit gender and ethnicity bias exposed by six popular open-source image generation models towards SE tasks. We obtain 20,160 images by feeding each model with three sets of prompts describing different software-related tasks: One set does not include any specification of the person performing the task, one set specifies that the person performing the task is a Software Engineer , and the last set explicitly request a fair representation of different genders and ethnicities. Next, we evaluate the gender and ethnicity disparities in the generated images. Results: The results indicate that all models exhibit a significant bias related to gender and ethnicity in SE tasks. Furthermore, we demonstrate that prompt engineering effectively reduces gender bias in only one of the six models; however, none of the models achieves fair representation with respect to ethnicity. Conclusion: The results of our analysis highlight serious concerns about the adoption of these models to generate content for SE tasks and open the field for future research on bias mitigation in this context.

Abstract Fairness is a critical concept in ethics and social domains, but it is also a challenging property to engineer in software systems. With the increasing use of machine learning in software systems, researchers have been developing techniques to assess the fairness of software systems automatically. Nonetheless, many of these techniques rely upon pre-established fairness definitions, metrics, and criteria, which may fail to encompass the wide-ranging needs and preferences of users and stakeholders. To overcome this limitation, we propose a novel approach, called MODNESS, that enables users to customize and define their fairness concepts using a dedicated modeling environment. Our approach guides the user through the definition of new fairness concepts also in emerging domains, and the specification and composition of metrics for its evaluation through a dedicated domain-specific language. Ultimately, MODNESS generates the source code to implement fair assessment based on these custom definitions. In addition, we elucidate the process we followed to collect and analyze relevant literature on fairness assessment in software engineering (SE). We compare MODNESS with the selected approaches and evaluate how they support the distinguishing features identified by our study. Our findings reveal that i) most of the current approaches do not support user-defined fairness concepts; ii) our approach can cover additional application domains not addressed by currently available tools, e.g., mitigating bias in recommender systems for software engineering and Arduino software component recommendations; iii) MODNESS demonstrates the capability to overcome the limitations of the only two other model-driven engineering-based approaches for fairness assessment.

A few years after their release, Large Language Models (LLMs)-based tools are becoming an essential component of software education, as calculators are used in math courses. When learning software engineering (SE), the challenge is the extent to which LLMs are suitable and easy to use for different software development tasks. In this paper, we report the findings and lessons learned from using LLM-based tools-ChatGPT in particular-in five SE courses from four universities. After instructing students on the LLM potentials in SE and about prompting strategies, we ask participants to complete a survey and be involved in semi-structured interviews. The collected results report (i) indications about the usefulness of the LLM for different tasks, (ii) challenges to prompt the LLM, i.e., interact with it, (iii) challenges to adapt the generated artifacts to their own needs, and (iv) wishes about some valuable features students would like to see in LLM-based tools. Although results vary among different courses, also because of students' seniority and course goals, the perceived usefulness is greater for lowlevel phases (e.g., coding or debugging/fault localization) than for analysis and design phases. Interaction and code adaptation challenges vary among tasks and are mostly related to the need for task-specific prompts, as well as better specification of the development context.

Code obfuscation is the conversion of original source code into a functionally equivalent but less readable form, aiming to prevent reverse engineering and intellectual property theft. This is a challenging task since it is crucial to maintain functional correctness of the code while substantially disguising the input code. The recent development of large language models (LLMs) paves the way for practical applications in different domains, including software engineering. This work performs an empirical study on the ability of LLMs to obfuscate Python source code and introduces a metric (i.e., semantic elasticity) to measure the quality degree of obfuscated code. We experimented with 3 leading LLMs, i.e., Claude-3.5-Sonnet, Gemini-1.5, GPT-4-Turbo across 30 Python functions from diverse computational domains. Our findings reveal GPT-4-Turbo’s remarkable effectiveness with few-shot prompting (81% pass rate versus 29% standard prompting), significantly outperforming both Gemini-1.5 (39%) and Claude-3.5-Sonnet (30%). Notably, we discovered a counter-intuitive “obfuscation by simplification” phenomenon where models consistently reduce rather than increase cyclomatic complexity. This study provides a methodological framework for evaluating AI-driven obfuscation while highlighting promising directions for leveraging LLMs in software security.

Language models have proven successful across a wide range of software engineering tasks, but their significant computational costs often hinder their practical adoption. To address this challenge, researchers have begun applying various compression strategies to improve the efficiency of language models for code. These strategies aim to optimize inference latency and memory usage, though often at the cost of reduced model effectiveness. However, there is still a significant gap in understanding how these strategies influence the efficiency and effectiveness of language models for code. Here, we empirically investigate the impact of three well-known compression strategies - knowledge distillation, quantization, and pruning - across three different classes of software engineering tasks: vulnerability detection, code summarization, and code search. Our findings reveal that the impact of these strategies varies greatly depending on the task and the specific compression method employed. Practitioners and researchers can use these insights to make informed decisions when selecting the most appropriate compression strategy, balancing both efficiency and effectiveness based on their specific needs.

The rise of AI software has made fairness auditing essential, particularly where biased decisions have serious impacts. This entails identifying sensitive variables and calculating fairness metrics based on predictions from a baseline model. Since model training is computationally intensive, recent research focuses on early bias assessment to detect bias before extensive training starts. This paper presents an empirical study to evaluate how dataset statistics, named bias symptoms , can assist in the early identification of variables that may lead to bias in the system. The aim of this study is to avoid training a machine learning model before assessing - and, in case, mitigating - its bias, thus increasing the sustainability of the development process. We first identify a bias symptoms dataset, employing 24 datasets from diverse application domains commonly used in fairness auditing. Through extensive empirical analysis, we investigate the ability of these bias symptoms to predict variables associated with bias under three fairness definitions. Our results demonstrate that bias symptoms are effective in supporting early predictions of bias-inducing variables under specific fairness definitions. These findings offer valuable insights for practitioners and researchers, encouraging further exploration in developing methods for proactive bias mitigation involving bias symptoms.

Machine learning (ML) is essential in modern technology, driving complex data-driven decisions. By 2025, daily data generation will exceed 463 exabytes, increasing ML’s influence and ethical risks of data exploitation and discrimination. The European Union’s Artificial Intelligence Act highlights the need for ethical AI solutions.

Gender gap in education has gained considerable attention in recent years, as it carries profound implications for the academic community. However, while the problem has been tackled from a student perspective, research is still lacking from an academic point of view. In this work, our main objective is to address this unexplored area by shedding light on the intricate dynamics of gender gap within the Software Engineering (SE) community. To this aim, we first review how the problem of gender gap in the SE community and in academia has been addressed by the literature so far. Results show that men in SE build more tightly-knit clusters but less global co-authorship relations than women, but the networks do not exhibit homophily. Concerning academic promotions, the Software Engineering community presents a higher bias in promotions to Associate Professors and a smaller bias in promotions to Full Professors than the overall Informatics community.

The continuous software engineering paradigm is gaining popularity in modern development practices, where the interleaving of design and runtime activities is induced by the continuous evolution of software systems. In this context, performance assessment is not easy, but recent studies have shown that architectural models evolving with the software can support this goal. In this paper, we present a mapping study aimed at classifying existing scientific contributions that deal with the architectural support for performance-targeted continuous software engineering. We have applied the systematic mapping methodology to an initial set of 215 potentially relevant papers and selected 66 primary studies that we have analyzed to characterize and classify the current state of research. This classification helps to focus on the main aspects that are being considered in this domain and, mostly, on the emerging findings and implications for future research. Editor’s note: Open Science material was validated by the Journal of Systems and Software Open Science Board. (see [https://www.sciencedirect.com/science/article/pii/S0164121221002168] for an example for where to place the statement and how to format it).

Nowadays, physicians have at their hands a huge amount of data produced by a large set of diagnostic and instrumental tests integrated with data obtained by high-throughput technologies. If such data were opportunely linked and analysed, they might be used to strengthen predictions, so that to improve the prevention and the time-to-diagnosis, reduce the costs of the health system, and bring out hidden knowledge. Machine learning is the principal technique used nowadays to leverage data and gain useful information. However, it has led to various challenges, such as improving the interpretability and explainability of the employed predictive models and integrating expert knowledge into the final system. Solving those challenges is of paramount importance to enhance the trust of both clinicians and patients in the system predictions. To solve the aforementioned issues, in this paper we propose a software workflow able to cope with the trustworthiness aspects of machine learning models and considering a multitude of heterogeneous data and models.

One of the first step in RNA-Sequencing (RNA-Seq) data analysis consists of aligning (Next Generation Sequencing) reads to a reference genome. In literature, there are several tools implemented by practitioners and researchers for the alignment step. However, two tools are the de-facto-standard used by bioinformatics researchers in their pipelines: HISAT (version 2) and STAR (version 2). The aim of this study is to determine the impact of the alignment tool on the RNA -Seq analysis in terms of biological relevance of the results and computational time. The two implemented pipelines return different results on the biological side. This is due to assumptions the used tools made and to the specific characteristics of the underlying (statistical) models. The study provides valuable insights for researchers interested in optimizing their RNA-Seq pipelines and making informed decisions about which pipeline to use. As lesson learned, we suggest bioinformatics researchers to use more pipelines when make experiments to reduce the prediction errors induced by assumption of a specific tool or method.

Abstract ML systems have become an essential tool for experts of many domains, data scientists and researchers, allowing them to find answers to many complex business questions starting from raw datasets. Nevertheless, the development of ML systems able to satisfy the stakeholders’ needs requires an appropriate amount of knowledge about the ML domain. Over the years, several solutions have been proposed to automate the development of ML systems. However, an approach taking into account the new quality concerns needed by ML systems (like fairness, interpretability, privacy, and others) is still missing. In this paper, we propose a new engineering approach for the quality-based development of ML systems by realizing a workflow formalized as a Software Product Line through Extended Feature Models to generate an ML System satisfying the required quality constraints. The proposed approach leverages an experimental environment that applies all the settings to enhance a given Quality Attribute, and selects the best one. The experimental environment is general and can be used for future quality methods’ evaluations. Finally, we demonstrate the usefulness of our approach in the context of multi-class classification problem and fairness quality attribute.

Nowadays assuring that search and recommendation systems are fair and do not apply discrimination among any kind of population has become of paramount importance. This is also highlighted by some of the sustainable development goals proposed by the United Nations. Those systems typically rely on machine learning algorithms that solve the classification task. Although the problem of fairness has been widely addressed in binary classification, unfortunately, the fairness of multi-class classification problem needs to be further investigated lacking well-established solutions. For the aforementioned reasons, in this paper, we present the Debiaser for Multiple Variables (DEMV), an approach able to mitigate unbalanced groups bias (i.e., bias caused by an unequal distribution of instances in the population) in both binary and multi-class classification problems with multiple sensitive variables. The proposed method is compared, under several conditions, with a set of well-established baselines using different categories of classifiers. At first we conduct a specific study to understand which is the best generation strategies and their impact on DEMV’s ability to improve fairness. Then, we evaluate our method on a heterogeneous set of datasets and we show how it overcomes the established algorithms of the literature in the multi-class classification setting and in the binary classification setting when more than two sensitive variables are involved. Finally, based on the conducted experiments, we discuss strengths and weaknesses of our method and of the other baselines.

Natural disasters always have several effects on human lives. It is always challenging for governments to tackle these incidents and to rebuild the economic, social and physical infrastructures and facilities with the available resources (mainly budget and time). The governments always define plans and policies in accordance with the law and political strategies that should maximize social benefits. The severity of damage and the huge resources needed to bring back life to normality make such reconstruction challenging. This article presents an approach to decision-support system by using deep reinforcement learning technique for the planning of post-disaster city reconstruction by considering available resources, meeting the needs of the broad community stakeholders (like citizens’ social benefits and politicians’ priorities) and keeping in consideration city’s structural constraints (like dependencies among roads and buildings). The proposed approach post disaster Rebuilding Plan Provider (pd-RPP) is generic, can determine a set of alternative plans for local administrators who select the ideal one to implement and it can be applied to areas of any extension. We show the proposed approach on a district of Sulmona city in Italy.

Context awareness is a first-class attribute of today software systems. Indeed, many applications need to be aware of their context in order to adapt their structure and behavior for offering the best quality of service even in case the software and hardware resources are limited. Modeling the context, its evolution, and its influence on the services provided by (possibly resource constrained) applications are becoming primary activities throughout the whole software life cycle, although it is still difficult to capture the multidimensional nature of context. We propose a framework for modeling and reasoning on the context and its evolution along multiple dimensions. Our approach enables (1) the representation of dependencies among heterogeneous context attributes through a formally defined semantics for attribute composition and (2) the stochastic analysis of context evolution. As a result, context can be part of a model-based software development process, and multidimensional context analysis can be used for different purposes, such as non-functional analysis. We demonstrate how certain types of analysis, not feasible with context-agnostic approaches, are enabled in our framework by explicitly representing the interplay between context evolution and non-functional attributes. Such analyses allow the identification of critical aspects or design errors that may not emerge without jointly taking into account multiple context attributes. The framework is shown at work on a case study in the eHealth domain.

BACKGROUND: Hepatocellular carcinoma (HCC) is an aggressive epithelial tumor which shows very poor prognosis and high rate of recurrence, representing an urgent problem for public healthcare. MicroRNAs (miRNAs/miRs) are a class of small, non-coding RNAs that attract great attention because of their role in regulation of processes such as cellular growth, proliferation, apoptosis. Because of the thousands of potential interactions between a single miR and target mRNAs, bioinformatics prediction tools are very useful to facilitate the task for individuating and selecting putative target genes. In this study, we present a chemically-induced HCC mouse model to identify differential expression of miRNAs during the progression of the hepatic injury up to HCC onset. In addition, we describe an established bioinformatics approach to highlight putative target genes and protein interaction networks where they are involved. RESULTS: We describe four miRs (miR-125a-5p, miR-27a, miR-182, miR-193b) which showed to be differentially expressed in the chemically-induced HCC mouse model. The miRs were subjected to four of the most used predictions tools and 15 predicted target genes were identified. The expression of one (ANK3) among the 15 predicted targets was further validated by immunoblotting. Then, enrichment annotation analysis was performed revealing significant clusters, including some playing a role in ion transporter activity, regulation of receptor protein serine/threonine kinase signaling pathway, protein import into nucleus, regulation of intracellular protein transport, regulation of cell adhesion, growth factor binding, and regulation of TGF-beta/SMAD signaling pathway. A network construction was created and links between the selected miRs, the predicted targets as well as the possible interactions among them and other proteins were built up. CONCLUSIONS: In this study, we combined miRNA expression analysis, obtained by an in vivo HCC mouse model, with a bioinformatics-based workflow. New genes, pathways and protein interactions, putatively involved in HCC initiation and progression, were identified and explored.

The problem of interpreting the results of performance analysis is quite critical, mostly because the analysis results (i.e. mean values, variances, and probability distributions) are hard to transform into feedback for software engineers that allows to remove performance problems. Approaches aimed at identifying and removing the causes of poor performance in software systems commonly fall in two categories: (i) bottleneck analysis, aimed at identifying overloaded software components and/or hardware resources that affect the whole system performance, and (ii) performance antipatterns, aimed at detecting and removing common design mistakes that notably induce performance degradation.

Wireless Sensor Networks are becoming one of the most successful choices for the development and deployment of applications in a range of scenarios, from intelligent homes to environment monitoring. Nowadays, there is a growing demand for programming large-scale wireless sensor networks. New programming paradigms should ease the task of building WSN applications that adapt at run-time to changes in the context, in the available resources, and also in user requirements. In this paper we describe PROTEUS, a platform to manage adaptation and reconfiguration, with the aim of supporting the development of WSN applications. After introducing PROTEUS, we show how it can be used to program a dynamic clustering algorithm, where clusters are created and destroyed at runtime, and nodes need to adapt and reconfigure accordingly. We provide a prototype implementation using TinyOS. Some remarks on the work are also presented.

Wireless Sensor Networks are becoming one of the most successful choices for the development and deployment of a wide range of applications, from intelligent homes to environment monitoring. Nowadays, there is a growing demand for fast development of WSN applications that adapt at run-time to changes in the context, in the available resources, and also in user requirements. In this paper we present a model-driven approach that permits to model and automatically generate Agilla Agents. We decide to target Agilla since it is an agent based platform that allows to manage adaptation without service interruptions by means of agents substitution. The proposed approach considers UML (Unified Modeling Language) as modeling language, and consists of a UML Profile to model Agilla agents and of a Model-to-Code transformation that generates Agilla code from the presented UML modeling framework.

We present a research roadmap that defines an enhanced model-driven software engineering approach focused on non-functional properties models. Currently, we have implemented two sub-processes of this roadmap: Property Modeling and Monitoring. We provide a property-driven approach to runtime monitoring based on a comprehensive Property Meta- Model (PMM) and on a generic configurable event-based monitoring infrastructure.

This paper presents a project that has been recently funded by means of the ERC Starting Grant (http://erc.europa.eu/). The main goal of the project is to develop a new generarion of wireless sensor networks infrastructure to support innovative services for ubiquitous sensing and video. The work is in a very preliminary phase so the paper presents a general overview and the main challenges.

The problem of interpreting the results of performance analysis is quite critical in the software performance domain. Mean values, variances, probability distributions are hard to interpret for providing feedback to software architects. Instead, what architects expect are solutions to performance problems, possibly in the form of architectural alternatives (e. g. split a software component in two components and re-deploy one of them). In a software performance engineering approach this path from analysis results to software alternatives still lacks of automation and is based on the skills and experience of analysts. In this paper we propose an automated approach for the performance feedback generation process based on performance antipatterns. To this aim, we model performance antipatterns as logical predicates and we provide a java engine, based on such predicates, that is able to detect performance antipatterns in an XML representation of the software system. Finally, we show the approach at work on a simple case study.

Provides notice of upcoming conference events of interest to practitioners and researchers.

Software models are, in most cases, considered as functional abstractions of systems. They represent the backbone of transformational processes aimed at code generation. On the other end, modeling is a traditional activity in the field of non-functional validation of software/hardware systems, although non-functional models found on different notations (such as Petri Nets) and embed additional information (such as the operational profile) with respect to software models. In this paper we widen the scope of model-driven architecture by introducing a Non-Functional-MDA framework that, beside the typical model transformations for code generation, embeds new types of model transformations that allow to generate non-functional models. For an uniform integration of these practices, we define Platform Independent/Specific Models in the non-functional domain.

It is widely recognized that in order to make performance validation an integrated activity along the software lifecycle it is crucial to be supported from automated approaches. Easiness to annotate software models with performance parameters (e.g. the operational profile) and automated translations of the annotated models into "ready-to-validate" models are the key challenges in this direction. Several methodologies have been introduced in the last few years to address these challenges. The tutorial introduces the attendance to the main methodologies for annotating and transforming software models into performance models.

Early performance analysis based on queueing network models (QNM) has been often proposed to support software designers during the software development process. These approaches aim at addressing performance issues as early as possible in order to reduce design failures. All of them try to adapt to software systems the well-known system performance analysis methodology. This implies that they assume at design time the availability of information about the hardware platform the software runs on. In recent years, we have proposed a methodology that allows quantitative reasoning on software aspects without considering hardware aspects. In this work, we extend our methodology to encompass a compositional approach to performance analysis of software architecture described by means of UML 2.0 diagrams. The main improvements include the characterization of architectural patterns and of their corresponding QNM pattern; the use of multi-chain queueing network as system target model and the identification of the information needed to parameterize the system model.

Over the last decade, a lot of research has been directed toward integrating performance analysis into the software development process. Traditional software development methods focus on software correctness, introducing performance issues later in the development process. This approach does not take into account the fact that performance problems may require considerable changes in design, for example, at the software architecture level, or even worse at the requirement analysis level. Several approaches were proposed in order to address early software performance analysis. Although some of them have been successfully applied, we are still far from seeing performance analysis integrated into ordinary software development. In this paper, we present a comprehensive review of recent research in the field of model-based performance prediction at software development time in order to assess the maturity of the field and point out promising research directions.