Search

[Context and motivation] User reviews published in mobile app repositories are essential for understanding user satisfaction and engagement within a specific market segment. [Question/problem] Manual analysis of reviews is impractical due to the large data volume, and automated analysis faces challenges like data synthesis and reporting. This complicates the task for app providers in identifying patterns and significant events, especially in assessing the influence of competitor apps. Furthermore, review-based research is mostly limited to a single app or a single app provider, excluding potential competition analysis. Consequently, there is an open research challenge in leveraging user reviews to support cross-app analysis within a specific market segment. [Principal ideas/results] Following a case-study research method in the microblogging app market, we introduce an automatic, novel approach to support mobile app market analysis. Our approach leverages quantitative metrics and event detection techniques based on newly published user reviews. Significant events are proactively identified and summarized by comparing metric deviations with historical baseline indicators within the lifecycle of a mobile app. [Contribution] Results from our case study show empirical evidence of the detection of relevant events within the selected market segment, including software- or release-based events, contextual events and the emergence of new competitors., With the support from the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund. This paper has been funded by the Spanish Ministerio de Ciencia e Innovación under project/funding scheme PID2020-117191RBI00/AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

(English) In the last two decades, the use of web Application Programming Interfaces (APIs) has grown exponentially, providing both consumers (software developers) and providers (companies and institutions that expose their organizational data or services) with numerous advantages and facilities. Web APIs allow developers to easily integrate existing services or data into their applications, reducing the time and cost required to build new software. At the same time, by making available their web APIs, providers can increase their customer reach or create a new revenue stream by monetizing the web API. Even though consumers expect web APIs to be steady and well-established, web APIs are prone to continuous changes, evolving several times through their lifecycle. As the use of web APIs continues to grow, the need to evolve them to meet the changing needs of consumers becomes more challenging. However, consumers do not always welcome web API evolution, as the changes might be too frequent or not relevant from their point of view. This discontentment becomes even stronger for the fact that web APIs are exposed over the Internet, meaning that consumers may be forced to upgrade to new web API versions (if providers decide to discontinue the former ones). Knowing the impact changes have on consumers, providers have to strike a balance between not imposing unexpected, frequent changes and providing an up-to-date, maintainable, bug-free web API, that fulfills consumers’ needs. The aim of this research work is to develop a semi-automatic method that enables web API providers to prescribe changes based on their consumers' behavior, as recorded in web API usage logs. We take the position that web API evolution should be mainly usage-based, i.e., the way consumers use web APIs should be one of the main drivers of web API changes. We start with the exploration of web API evolution, by seeing the process from the points of view of both providers and consumers. We identify and cla, (Català) En les últimes dues dècades, l'ús d'Aplicacions de Programació d'Interfícies (APIs) web ha crescut exponencialment, oferint nombroses avantatges i facilitats tant als consumidors (desenvolupadors de programari) com als proveïdors (empreses i institucions que exposen les seves dades o serveis organitzatius). Les APIs web permeten als desenvolupadors integrar fàcilment serveis o dades existents en les seves aplicacions, reduint el temps i el cost necessaris per construir nous programaris. Al mateix temps, al posar a disposició la seva API web, els proveïdors poden augmentar el nombre de clients de la seva marca o crear una nova font de guanys monetitzant la seva API web. Tot i que els consumidors esperen que les APIs web siguin estables i ben establertes, les APIs web són propenses a canvis contínues, evolucionant diverses vegades al llarg del seu cicle de vida. A mesura que l'ús de les APIs web continua creixent, la necessitat de fer-les evolucionar per satisfer les necessitats canviant dels consumidors esdevé més desafiadora. No obstant això, els consumidors no sempre acullen amb satisfacció l'evolució de les APIs web, ja que els canvis poden ser massa freqüents o no rellevants des del seu punt de vista. Aquesta descontentament esdevé encara més fort pel fet que les APIs web estan exposades a través d'Internet, el que significa que els consumidors poden ser obligats a actualitzar-se a noves versions de la API web (si els proveïdors decideixen discontinuar les antigues). Coneixent l'impacte que els canvis tenen en els consumidors, els proveïdors han de trobar un equilibri entre no imposar canvis inesperats i freqüents i proporcionar una API web actualitzada, mantenible i sense errors, que satisfaci les necessitats dels consumidors. L'objectiu d'aquest treball de recerca és desenvolupar un mètode semi-automàtic que permeti als proveïdors de web API prescriure canvis basats en el comportament dels seus consumidors, tal com es registra en els logs d'ús de la we, Postprint (published version)

Agile software development (ASD) aims to deliver high-quality software. However, software quality often receives less attention in ASD due to focus on functionality. Elicit requirements for, design, validate, implement, and evaluate an evidence-based quality-aware ASD process that organizations can use for better integration of quality management. Using design science research, we collaborated for over 1.5 years with four organizations applying ASD. Involving a total of 35 practitioners, we designed new evidence-based quality-aware ASD processes, which were validated through simulated scenarios, implemented, and evaluated by the practitioners. A general ASD process model, derived from the companyspecific models, was created; it shows activities, artifacts, roles, and development phases where evidence can be used to improve software quality. The new process model contributed to higher quality awareness, improved bottleneck identification, release readiness, and quality issues prioritization in the studied companies. We also learned that companies must be mindful of the rules that determine the use of evidence. Integrating the use of quality related evidence into the ASD process can make it quality-aware, support software quality management, and enable timely reaction through fast feedback loops. The evidence-based quality-aware ASD process is applicable in any Scrum-like ASD process. Applicability in other forms of ASD, however, requires further research., This work was supported by the Q-Rapids project though the European Union’s Horizon 2020 Research and Innovation Program under Grant 732253., Peer Reviewed, Postprint (published version)

Con las complejas interfaces de usuario de las aplicaciones modernas, es una práctica muy común y buena probar la interfaz de usuario en cada actualización para asegurar que la aplicación funcione como se pretende. En GTD esto es muy común y los tests normalmente se ejecutan manualmente por un desarrollador siguiendo un conjunto de instrucciones, que piden al desarrollador verificar el estado de la interfaz de usuario frente a descripciones o imágenes de lo que debería suceder. El objetivo de este proyecto es desarrollar una solución para automatizar este proceso de prueba tanto como sea posible, para reducir el tiempo actual invertido en probar la interfaz de usuario y aumentar la frecuencia con la que se pueden realizar los tests., With modern applications having complex user interfaces, it is a very common and good practice to test the user interface every release to ensure the application works as intended. In GTD this is very common and the tests are usually run manually by a developer, following a set of instructions steps that ask the developer to verify the state of the UI against descriptions or images of what should happen. The goal of this project is to develop a solution to automate this testing process as much as possible, to both reduce the current time spent testing the user interface and increase the frequency tests can be run.

Este Trabajo de Final de Grado, desarrollado en el marco del grado en Ingeniería Informática en la Facultad de Informática de Barcelona (FIB), tiene como objetivo el desarrollo de un servicio capaz de reproducir contenido multimedia en varios dispositivos web de manera simultánea, así como el diseño y la construcción de la infraestructura en la que se aloja dicha página web. La reproducción de contenido ya era funcional en televisores, pero este proyecto afronta el reto de llevar la reproducción a navegadores, además de poder sincronizarlos entre ellos y con los televisores desde la app móvil. Esta nueva funcionalidad servirá para brindar una experiencia más satisfactoria a los clientes de la aplicación y abrirá las puertas a nuevas colaboraciones con partners., This Final Degree Project, developed within the framework of the degree in Computer Engineering at the Barcelona School of Informatics (FIB), aims to develop a service capable of playing multimedia content on various web devices simultaneously, as well as designing and building the infrastructure where this web page is hosted. Content playback was already functional on televisions, but this project tackles the challenge of bringing playback to browsers, as well as synchronizing them with each other and with the televisions from the mobile app. This new functionality will provide a more satisfying experience for the application's clients and open doors to new collaborations with partners.

The Integrated Gamified Learning Dashboard Environment (GLiDE) project aims at fostering student engagement, teamwork, and project performance in software engineering education. By integrating gamification elements and enhanced user experience in an existing Learning Dashboard, GLiDE offers a novel approach to learning in team-based software development projects. This paper discusses the motivation behind GLiDE, its current status, and the open research areas. This project offers an opportunity to gain valuable insights into the fields of educational technology and software engineering education., Funding Agency: Universitat Politècnica de Catalunya; Period: May 2023-July 2024; URL: https://gessi.upc.edu/en/projects/glide-gamified-learning-dashboard-environment., Peer Reviewed, Postprint (published version)

Resolved, 2023/2024, 1r quadrimestre

Background: The usage of complex artificial intelligence (AI) models demands expensive computational resources. While currently, available high-performance computing environments can support such complexity, the deployment of AI models in mobile devices, which is an increasing trend, is challenging. Environments with low computational resources imply limitations in the design decisions during the AI-enabled software engineering lifecycle that balance the trade-off between the accuracy and the complexity of the mobile applications. Objective: Our objective is to systematically assess the trade-off between accuracy and complexity when deploying complex AI models (e.g. neural networks) to mobile devices in pursuit of greener AI solutions. We aim to cover (i) the impact of the design decisions on the achievement of high-accuracy and low resource-consumption implementations; and (ii) the validation of profiling tools for systematically promoting greener AI. Method: We implement neural networks in mobile applications to solve multiple image and text classification problems on a variety of benchmark datasets. We then profile and model the accuracy, storage weight, and time of CPU usage of the AI-enabled applications in operation with respect to their design decisions. Finally, we provide an open-source data repository following the EMSE open science practices and containing all the experimentation, analysis, and reports in our study. Results: We find that the number of parameters in the AI models makes the time of CPU usage scale exponentially in convolutional neural networks and logarithmically in fully-connected layers. We also see the storage weight scales linearly with the number of parameters, while the accuracy does not. For this reason, we argue that a good practice for practitioners is to start small and only increase the size of the AI models when their accuracy is low. We also find that Residual Networks (ResNets) and Transformers have a higher baseline cost in t, This work has been partially supported by the GAISSA project (TED2021-130923B-I00, which is funded by MCIN/AEI/10.13039/501100011033 and by the European Union “NextGenerationEU”/PRTR); and, by the “Beatriz Galindo” Spanish Program (BEAGAL18/00064)., Peer Reviewed, Postprint (author's final draft)

Nowadays, AI-based systems have achieved out-standing results and have outperformed humans in different domains. However, the processes of training AI models and inferring from them require high computational resources, which pose a significant challenge in the current energy efficiency societal demand. To cope with this challenge, this research project paper describes the main vision, goals, and expected outcomes of the GAISSA project. The GAISSA project aims at providing data scientists and software engineers tool-supported, architecture-centric methods for the modelling and development of green AI-based systems. Although the project is in an initial stage,we describe the current research results, which illustrate the potential to achieve GAISSA objectives., This paper has been funded by the GAISSA Spanish research project (ref. TED2021-130923B-I00; MCIN/AEI/10.13039/501100011033)., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Objectius de Desenvolupament Sostenible::13 - Acció per al Clima, Postprint (author's final draft)

Actualment, es disposa del Learning Dashboard, un tauler de control que permet el seguiment del desenvolupament de projectes software realitzats per estudiants. Per les visualitzacions d'aquests, s'ha desenvolupat una interfície web que permet realitzar el seguiment a partir de diverses mètriques o indicadors que marquen el progrés. En aquest projecte, es contempla la possibilitat de millorar les visualitzacions ja existents mitjançant la realització d'una nova aplicació web desenvolupada en Angular i NestJS., Currently the Learning Dashboard is available, a control panel that enables monitoring of development of software projects carried out by student. For the visualizations of these, a web interface has been developed that allows monitoring based on various metrics or indicators that mark the progress. In this project, we are considering the possibility of improving the existing visualizations through the creation of a new web application developed in Angular and NestJS.

Task-oriented chatbots relying on a knowledge base for domain-specific content exploitation have been largely addressed in research and industry applications. Despite this, multiple challenges remain to be fully conquered, including adaptive knowledge mechanisms, personalization for user-specific demands, and composite intent resolution. To address these challenges, in this paper, we present a work-in-progress summary of a task-oriented, knowledge-based chatbot in the field of mobile software ecosystems. The chatbot is designed to assist users in the combined use of multiple features from different applications. The proposed knowledge base and the machine learning pipeline supporting the chatbot technical core are designed to: (i) effectively use user context, (ii) process runtime feedback, (iii) use user historical data, and (iv) automatically infer slot values and dependent actions. With this report, we expect to lay the groundwork for future development stages and user validation studies., With the support from the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund. This paper has been funded by the Spanish Ministerio de Ciencia e Innovación under project / funding scheme PID2020-117191RB-I00 / AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

Background: The rise of machine learning (ML) systems has exacerbated their carbon footprint due to increased capabilities and model sizes. However, there is scarce knowledge on how the carbon footprint of ML models is actually measured, reported, and evaluated. Aims: This paper analyzes the measurement of the carbon footprint of 1,417 ML models and associated datasets on Hugging Face. Hugging Face is the most popular repository for pretrained ML models. We aim to provide insights and recommendations on how to report and optimize the carbon efficiency of ML models. Method: We conduct the first repository mining study on the Hugging Face Hub API on carbon emissions and answer two research questions: (1) how do ML model creators measure and report carbon emissions on Hugging Face Hub?, and (2) what aspects impact the carbon emissions of training ML models? Results: Key findings from the study include a stalled proportion of carbon emissions-reporting models, a slight decrease in reported carbon footprint on Hugging Face over the past 2 years, and a continued dominance of NLP as the main application domain reporting emissions. The study also uncovers correlations between carbon emissions and various attributes, such as model size, dataset size, ML application domains and performance metrics. Conclusions: The results emphasize the need for software measurements to improve energy reporting practices and the promotion of carbon-efficient model development within the Hugging Face community. To address this issue, we propose two classifications: one for categorizing models based on their carbon emission reporting practices and another for their carbon efficiency. With these classification proposals, we aim to encourage transparency and sustainable model development within the ML community., This work is partially supported by the project TED2021- 130923B-I00, funded by MCIN/AEI/10.13039/501100011033 and the European Union Next Generation EU/PRTR., Peer Reviewed, Postprint (author's final draft)

Resolved, 2023/2024, 1r quadrimestre

La proposta d'aquest treball és proporcionar als professors mètriques i indicadors que els ajudin a monitoritzar el desenvolupament de treballs software fet per estudiants gràcies a la informació que proporcionen les eines de gestió del procés del projecte software, versionat de codi i més, utilitzats pels alumnes. En aquesta tesi s'investigarà quines funcionalitats tenen els Learning Dashboards ja creats seguint una metodologia SLR per estudiar els articles científics i com veure quines millores es poden incorporar al Learning Dashboard creat per la UPC per les assignatures d'Aplicacions i Serveis Web (ASW) i Projecte d'Enginyeria de Software (PES) del Grau d'Enginyeria Informàtica per tal d'ajudar al professorat a monitoritzar el progrés i rendiment dels alumnes en el projecte de desenvolupament de software impartit en aquestes dues assignatures.

(English) This thesis presents the Tychonis framework, a solution to the limitations faced by existing software packages used in space missions to identify geometric events. The framework is designed to integrate with current and future mission software, providing users with the ability to extend opportunities and search algorithms without modifying the tools that use it. The framework is built on the principles of separation of concerns, extensibility, reusability, and independent verification and validation. It is provided as a software library, allowing missions to add their own data structures and constructs, promoting cross-mission reusability. The text also includes a case study of Tychonis' integration with the Science Opportunity Analyzer (SOA) software, demonstrating its ability to be extrapolated to other tools that need to search for geometric events. As a complement to the Tychonis framework, the thesis introduces two computer languages designed to make the process of modeling opportunities more accessible. Scientists, who may face difficulties with imperative programming languages or lack of available science planning tools, are the target audience for these languages. The readability and usability of the languages were evaluated through a comprehensive study involving a questionnaire with active exercises, statements with corresponding responses on a Likert scale, and open-ended questions to elicit qualitative responses. The results provide both relative and absolute quantification of the usability and readability of each language, as well as qualitative results to direct future language design decisions. Emphasizing the importance of utilizing proven software principles and good design choices in space missions helps reduce risk and cost. Tychonis and our research on accessible computer languages embrace this concept., (Català) Aquesta tesi presenta el framework Tychonis, una solució a les limitacions que enfronten els paquets de programari existents utilitzats en les missions espacials per identificar esdeveniments geomètrics. El framework està dissenyat per integrar-se amb el programari de missió actual i futur, proporcionant als usuaris la capacitat d'estendre les oportunitats i els algoritmes de cerca sense modificar les eines que el fan servir. El framework es basa en els principis de separació d’interessos, extensibilitat, reutilització, i verificació i validació independents. Es proporciona com a biblioteca de programari, permetent que les missions afegiu les seves pròpies estructures de dades i construccions, promocionant la reutilització entre missions. El text també inclou un estudi de cas de la integració de Tychonis amb el programari Science Opportunity Analyzer (SOA), demostrant la seva capacitat per ser extrapolada a altres eines que necessiten buscar esdeveniments geomètrics. Com a complement a Tychonis, la tesi presenta dos llenguatges de programació dissenyats per fer més accessible el procés de modelització d'oportunitats. Els científics, que poden trobar dificultats amb els llenguatges de programació imperatius o la falta d'eines de planificació científica disponibles, són el públic objectiu d'aquests llenguatges. La llegibilitat i la usabilitat dels llenguatges s’han avaluat a través d'un estudi exhaustiu que inclou un qüestionari amb exercicis actius, afirmacions amb respostes en una escala Likert, i preguntes obertes per obtenir respostes qualitatives. Els resultats proporcionen quantificació relativa i absoluta de la usabilitat i la llegibilitat de cada llenguatge, així com resultats qualitatius per dirigir les decisions de disseny de futurs llenguatges. Posar èmfasi en la importància d'utilitzar bons principis de programació i bones eleccions de disseny en les missions espacials ajuda a reduir el risc i el cost Tychonis i la nostra investigació sobre llengu, Postprint (published version)

Current research in the computer vision field mainly focuses on improving Deep Learning (DL) correctness and inference time performance. However, there is still little work on the huge carbon footprint that has training DL models. This study aims to analyze the impact of the model architecture and training environment when training greener computer vision models. We divide this goal into two research questions. First, we analyze the effects of model architecture on achieving greener models while keeping correctness at optimal levels. Second, we study the influence of the training environment on producing greener models. To investigate these relationships, we collect multiple metrics related to energy efficiency and model correctness during the models’ training. Then, we outline the trade-offs between the measured energy efficiency and the models’ correctness regarding model architecture, and their relationship with the training environment. We conduct this research in the context of a computer vision system for image classification. In conclusion, we show that selecting the proper model architecture and training environment can reduce energy consumption dramatically (up to 98.83%) at the cost of negligible decreases in correctness. Also, we find evidence that GPUs should scale with the models’ computational complexity for better energy efficiency., This work is part of the GAISSA project (TED2021-130923B-I00), which is funded by MCIN/AEI/10.13039/501100011033 and by the European Union "NextGenerationEU”/PRTR., Peer Reviewed, Postprint (author's final draft)

El Learning Dashboard és un software implementat en el departament ESSI que mostra el valor de certes mètriques i indicadors relacionats amb el treball dels estudiants durant el desenvolupament en equip de projectes de software. Algunes d'aquestes mètriques i indicadors es basen en l'anàlisi de dades recollides de l'eina Taiga per a la gestió de projectes software. L'objectiu d'aquest TFG és incorporar aquestes mètriques i indicadors directament a Taiga enlloc de tenir-les externes.

El procés previ a la publicació dels articles d'una conferència científica implica la col·laboració de desenes de membres d'un comité que revisen, evaluen, validen i comenten aquests articles. La tasca del director, o "chair", d'assegurar la qualitat del procés de principi a fi no és senzilla. És per això que he desenvolupat WatChair, una servei que a partir d'un fitxer amb les dades de la conferència sintetitza unes mètriques per poder analitzar el seu estat. Aquestes es posen a la disposició de l'usuari mitjançant una API i una aplicació web que les presenta de forma més visual., Before an academic conference ends up publishing articles, there's a process involving dozens of members of a committee where they review, score, validate and discuss these works. The task of the steering committee chair is to ensure the quality throughout the whole process, and it's not an easy one. That's why we developed WatChair, a service that takes in a file with an academic conference data and creates a set of metrics to ease the analysis process. These are given to the user both through an API and a web application where they are presented in a visual manner.

En el ámbito académico, las asignaturas orientadas al desarrollo de proyectos presentan un gran reto, tanto para el profesorado como para el alumnado. Este reto consiste en poder cuantificar objetivamente el rendimiento que cada miembro del grupo está teniendo en el desarrollo del proyecto. Más allá del dato básico y utilizado por excelencia, que es el de las horas totales dedicadas, existen muchos otros indicadores que pueden mostrar el rendimiento real de cada integrante. Se ha observado que el alumnado utilizaba herramientas externas como SonarCloud y Google Sheets para poder autoevaluar el estado del código del proyecto y la dedicación por horas y actividad que dedicaba cada integrante al proyecto. El objetivo del Trabajo de Fin de Grado (TFG) es desarrollar dos conectores de datos a SonarCloud y Google Sheets, obtener con estos la información que ambas aplicaciones generan y traspasarla a ElasticSearch. Ello permitirá generar métricas de rendimiento homogéneas para todos los grupos de estudiantes de la misma asignatura. Como resultado de este proyecto, se implementaron ambos conectores. También la creación de un sistema de imputaciones horarias automatizado, en el que los requisitos de este fueron definidos en colaboración con el profesorado. Se hizo, además, un estudio del estado actual de la arquitectura de los conectores y propusieron mejoras para futuros evolutivos. Por último, se redactaron dos manuales de instrucciones para que el profesorado pueda utilizar las herramientas desarrolladas en este proyecto en el futuro. Estos resultados se han conseguido mediante el uso de metodologías ágiles que han permitido la constante adaptación del proyecto a las necesidades que han ido surgiendo por parte del profesorado. También estas metodologías han permitido crear un canal de comunicación constante con los directores de proyecto y profesores, lo que ha permitido que el desarrollo de este TFG haya estado siempre alineado con las expectativas del profesorado. Despu, In the academic field, project development subjects present a great challenge for both teachers and students. This challenge is that of not being able to objectively quantify the performance that each member of the group has in developing the project. Beyond the basic data used by excellence, that is to say the total number of hours spent, there are many other measures that can measure the actual performance of each person. Students took advantage of external tools such as SonarCloud and Google Sheets. They were used to self-assess the state of the project code and the time and dedication of each member treated in the project. This project focuses on the development of two data connectors for SonarCloud and Google Sheets. They are designed to obtain information generated by both applications and transfer them to ElasticSearch. In this way, uniform performance measures are generated for all groups of students in the same subject. As a result of this project, the two connectors have been implemented. Furthermore, the creation of an automated timeframe system, in which the requirements of this system have been defined in collaboration with teachers. A study of the current status of the connector architecture was also conducted and improvements proposed for future developments. Lastly, two instruction manuals have been written so that teachers can use the tools developed as part of this project in the future. These outcomes were achieved through agile methodologies. This has allowed the project to be continuously adapted to the needs that have emerged from the teachers. Moreover, these methodologies have made it possible to create a constant channel of communication with project directors and teachers, which has made the development of this "Trabajo de Fin de Grado (TFG)" aligned with teachers' expectations. After validating the implementation of both connectors and the timeframe system, we can conclude that its adoption in the Learning Dashboard can be beneficial for b

La naturalesa del processament de llengües naturals ha canviat dràsticament en els últims anys. La implementació de Large Language Models pre-entrenat en milers de dades sense etiquetar ha obert la porta a una nova capa de comprensió del processament de text. Això ha desplaçat la investigació a la zona per explotar aquests grans models per obtenir millors resultats per a les tasques més petites. D'aquesta manera, el processament de llengües naturals està adquirint una importància cada vegada major. Afinant els diferents models de llenguatge gran amb dades específiques de context i de tasques, aquests models ràpidament aprenen a seguir patrons i generalitzar-los a nous conceptes. Entenen el llenguatge natural en gran mesura i poden generar relacions en paraules, frases i paràgrafs. La sintonització fina neuronal s'ha convertit en una tasca cada vegada més important per simplificar l'ús de solucions d'aprenentatge automàtic amb pocs recursos. L'augment dels models de transformadors pre-entrenats per al processament del llenguatge natural ha complicat la selecció i l'experimentació d'aquests models, augmentant el temps de recerca i experimentació. Aquest estudi passa per l'estat actual de l'art dels models transformadors i intenta estudiar l'abast i l'aplicabilitat d'aquests models. A partir d'aquest treball inicial, el document produeix un gasoducte complet d'ajust fi del model que permet a l'usuari obtenir fàcilment un model llest per a utilitzar per a una tasca de llenguatge natural. Per provar aquest model, la canonada es prova i s'avalua per a l'extracció automàtica de característiques (és a dir, funcionalitats) des d'aplicacions mòbils utilitzant documents de llenguatge natural disponibles, com ara descripcions., The nature of Natural Language Processing has drastically changed in the past years. The implementation of Large Language Models pre-trained on thousands of unlabelled data has opened the door to a new layer of comprehension of text processing. This has shifted research in the area to exploit these large models to obtain better results for smaller tasks. In this way, fine-tuning Natural Language Processing is becoming increasingly important. By fine-tuning the different large language models with context and task-specific data, these models quickly learn to track patterns and generalize to new concepts. They understand natural language to a great extent and can generate relationships in words, phrases, and paragraphs. Fine Tuning has become an increasingly important task to simplify the use of machine learning solutions with low resources. The increase in pre-trained transformer models for Natural Language Processing has complicated the selection and experimentation of these models, increasing research and experimentation time. This study goes through the current state of the art of transformer models and attempts to study the scope and applicability of these models. From this initial work, the paper produces a compre- hensive pipeline of model fine-tuning that allows the user to easily obtain a ready-to-use model for a natural language task. To test this model, the pipeline is tested and evaluated for the automatic extraction of features (i.e. functionalities) from mobile applications using available natural language documents, such as descriptions.

Mobile app stores provide centralized access to a large data set of mobile app related natural language textual data, including developer’s documentation (e.g., descriptions, changelogs) and user-generated data (e.g., user reviews). Motivated by this context, multiple studies have focused on data-driven elicitation processes for the automatic extraction of the set of features exposed by a catalogue of applications and the inferred, extended knowledge that can be derived from this information. Moreover, with the emerging and generalization of large language models, traditional linguistic-based approaches can be significantly improved by the potential of the knowledge embedded in this kind of models. In this paper, we present TransFeatEx, a NLP-based feature extraction pipeline that combines the use of a RoBERTa-based model with the application of consolidated syntactic and semantic techniques. The pipeline is designed as a customizable, standalone service to be used either as a playground, experimentation tool or as a software component to be embedded into a third-party software system for batch-processing large document corpora. An example of a demo plan is showcased here: https://youtu.be/gfFyi_i_uTw, With the support from the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund. This paper has been funded by the Spanish Ministerio de Ciencia e Innovación under project / funding scheme PID2020-117191RB-I00 / AEI/10.13039/501100011033., Peer Reviewed, Postprint (published version)

Team-based software development projects (TBSDP) are a useful instrument to expose students to teamwork in an industry-like working context. However, TBSDP exposes students to a number of challenges. This paper has a twofold objective. First, understand the practices and challenges that students face in TBDSP in an Agile context. Second, investigate whether the use of a teaching domain-specific information system, a learning dashboard, could help them in improving these practices and facing these challenges. We conducted a multi-instrument exploratory study at the Polytechnical University of Catalunya. We gathered information about the progress of 39 students organised in 6 teams during one semester by mining two software repositories and conducting a number of questionnaires and interviews related both to working practices and to students’ perception of learning dashboard adoption. Results show that many students do not follow adequate practices for some TBSDP activities. On the other hand, metrics informing about the use of code repositories and task management were generally well understood and perceived as potentially useful by students when shown in a learning dashboard. We conclude that the adoption of learning dashboards is a viable approach to improve student practices in TBSDP, but it needs to be carefully considered which metrics provide the most value to face the identified challenges., This paper has been funded by MCIN/AEI/10.13039/ 501100011033 and NextGenerationEU/PRTR under project PDC2021-121195-I00., Peer Reviewed, Postprint (author's final draft)

Com a responsables o administradors d'una aplicació, poder disposar de certa informació sobre com els usuaris interaccionen amb aquesta pot resultar crucial a l'hora de detectar problemes o poder inferir millores. Per aquest motiu, l'objectiu d'aquest projecte és el d'implementar una aplicació que sigui capaç d'aconseguir aquesta informació sobre el Learning Dashboard, una aplicació gestionada pel departament GESSI de la Facultat d'Informàtica de Barcelona centrada en la monitorització d'informació potencialment rellevant pel desenvolupament de software en equip, i utilitzat actualment en dues assignatures del Grau en Enginyeria Informàtica de la facultat. A partir de la detecció de la mancança d'una eina d'aquestes caracterísitques, s'ha implementat una aplicació web, com a resultat final del projecte, que emmagatzema, tracta i presenta la informació sobre certs aspectes de l'ús que fan els usuaris del Learning Dashboard. L'objectiu principal de l'aplicació és facilitar informació potencialment rellevant a l'hora de prendre decisions amb relació al Learning Dashboard i el seu ús. Així mateix, aquest projecte també té com a objectiu indirecte millorar l'experiència d'usuari, tant d'alumnes com de professors, al Learning Dashboard. Això és perquè l'eina desenvolupada facilita el procés d'inferència de millores i solució de problemes d'usabilitat d'aquest, fet que inevitablement desemboca en una millora de l'experiència d'usuari i incentiva el seu ús. Aquests resultats s'han obtingut a través de l'execució d'una metodologia àgil de desenvolupament software, que ha permès una millor adaptació a imprevistos i dificultats que han sorgit en el procés de preparació, disseny i implementació del projecte. A més, els directors del projecte són administradors del Learning Dashboard i professors d'assignatures en què s'utilitza aquest, garantint així una bona alineació amb els requisits i objectius a assolir del treball., As managers or administrators of an application, being able to access certain information about how users interact with it can be crucial when detecting issues or inferring improvements. For this reason, the goal of this project is to implement an application that is capable of obtaining this information about the Learning Dashboard, an application managed by the GESSI department of the Faculty of Informatics in Barcelona, focused on monitoring potentially relevant information for team software development and currently used in two subjects of the Bachelor's Degree in Computer Engineering at the faculty. Based on the identification of the lack of a tool with these characteristics, a web application has been implemented as the final result of the project, which stores, processes, and displays information about certain aspects of the users' usage of the Learning Dashboard. The main objective of the application is to provide potentially relevant information when making decisions regarding the Learning Dashboard and its usage. Additionally, this project also has the indirect goal of improving the user experience for both students and teachers using the Learning Dashboard. This is because the developed tool facilitates the process of inferring improvements and solving usability issues, which inevitably leads to an enhanced user experience and encourages its usage. These results have been achieved through the execution of an agile software development methodology, which has allowed for better adaptation to unforeseen events and difficulties that have arisen during the preparation, design, and implementation of the project. Furthermore, the project directors are administrators of the Learning Dashboard and teachers of subjects in which it is used, ensuring a good alignment with the requirements and objectives established in the project.

Requirements specifcation is a core activity in the requirements engineering phase of a software development project. Researchers have contributed extensively to the feld of requirements specifcation, but the extent to which their proposals have been adopted in practice remains unclear. We gathered evidence about the state of practice in requirements specifcation by focussing on the artefacts used in this activity, the application of templates or guidelines, how requirements are structured in the specifcation document, what tools practitioners use to specify requirements, and what challenges they face. We conducted an interview-based survey study involving 24 practitioners from 12 diferent Swedish IT companies. We recorded the interviews and analysed these recordings, primarily by using qualitative methods. Natural language constitutes the main specifcation artefact but is usually accompanied by some other type of instrument. Most requirements specifcations use templates or guidelines, although they seldom follow any fxed standard. Requirements are always structured in the document according to the main functionalities of the system or to project areas or system parts. Diferent types of tools, including MS Ofce tools, are used, either individually or combined, in the compilation of requirements specifcations. We also note that challenges related to the use of natural language (dealing with ambiguity, inconsistency, and incompleteness) are the most frequent challenges that practitioners face in the compilation of requirements specifcations. These fndings are contextualized in terms of demographic factors related to the individual interviewees, the organization they are afliated with, and the project they selected to discuss during our interviews. A number of our fndings have been previously reported in related studies. These fndings show that, in spite of the large number of notations, models and tools proposed from academia for improving requirements specifcation, p, This paper has been funded by the Spanish Ministerio de Ciencia e Innovación under project/funding scheme PID2020-117191RB-I00/AEI/., Peer Reviewed, Postprint (published version)

Context and motivation: AI-based systems (i.e., systems integrating some AI model or component) are becoming pervasive in society. A number of characteristics of AI-based systems challenge classical requirements engineering (RE) and raise questions yet to be answered. Question: This vision paper inquires the role that RE should play in the development of AI-based systems with a focus on three areas: roles involved, requirements’ scope and non-functional requirements. Principal Ideas: The paper builds upon the vision that RE shall become the cornerstone in AI-based system development and proposes some initial ideas and roadmap for these three areas. Contribution: Our vision is a step towards clarifying the role of RE in the context of AI-based systems development. The different research lines outlined in the paper call for further research in this area., This paper is part of the project TED2021-130923B-I00, funded by MCIN/AEI/10.13039/501100011033 and the European Union “NextGenera-tionEU”/PRTR., Peer Reviewed, Postprint (author's final draft)

Recent surveys expose that the use of Open Source Software (OSS) is increasingly becoming a need for organizations in their development projects. However, deciding a proper OSS to be adopted or to contribute to its development is a complex and error-prone task. Analyzing the OSS ecosystem (OSSECO) health may help providing information about: (1) the OSS itself (number of commits, days after the last release, etc.); and (2) their main actors (number of contributors, partners, etc.). There exist proposals that go further and provide aggregated high-level indicators (e.g. visibility as an aggregation of number of community events, number of partners, and other metrics). Nevertheless, there is a lack of useful OSSECO analysis tools to ease the decision making on which OSSECO has the health required by a potential OSS adopter or contributor. In this work, we provide OSS-CARE (OSSeCo heAlthy monitoR and analysEr), an OSS-independent, fully automatic, and real-time framework to assess OSSECO’s health. OSS-CARE supports defining the ecosystem health objectives of potential OSS adopters, OSS contributors, and even OSS managers to inspect their provided health. These objectives are defined based on a well-established model characterizing health metrics that can be potentially aggregated by using a Bayesian network technique. Moreover, the integrated monitoring and analysis components perform an automated assessment of OSSECO’s health by checking the fulfillment of the required health objectives. Furthermore, the result is shown in an appealing dashboard that may ease the complex decision making of which OSS to choose., This work has been supported by the Spanish project PID2020-117191RB-I00 funded by MCIN/AEI/10.13039/501100011033; grants RTI2018-101204-B-C21, RTI2018-101204-B-C22, PID2021-126227NB-C21, PID2021-126227NB-C22 funded by MCIN/AEI/10.13039/501100011033 and ‘‘ERDF a way of making Europe’’; grants PYC20 RE 084 US, P18-FR-2895, US-1264651, US-1381595 funded by Junta de Andalucıa/ERDF, UE, Spain; and FPU19/00666 funded by MCIN/AEI/10.13039/501100011033 and by ‘‘ESF Investing in your future’’., Peer Reviewed, Postprint (author's final draft)

The use of natural language interfaces in the field of human-computer interaction is undergoing intense study through dedicated scientific and industrial research. The latest contributions in the field, including deep learning approaches like recurrent neural networks, the potential of context-aware strategies and user-centred design approaches, have brought back the attention of the community to software-based dialogue systems, generally known as conversational agents or chatbots. Nonetheless, and given the novelty of the field, a generic, context-independent overview on the current state of research of conversational agents covering all research perspectives involved is missing. Motivated by this context, this paper reports a survey of the current state of research of conversational agents through a systematic literature review of secondary studies. The conducted research is designed to develop an exhaustive perspective through a clear presentation of the aggregated knowledge published by recent literature within a variety of domains, research focuses and contexts. As a result, this research proposes a holistic taxonomy of the different dimensions involved in the conversational agents’ field, which is expected to help researchers and to lay the groundwork for future research in the field of natural language interfaces., With the support from the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund. The corresponding author gratefully acknowledges the Universitat Politècnica de Catalunya and Banco Santander for the inancial support of his predoctoral grant FPI-UPC. This paper has been funded by the Spanish Ministerio de Ciencia e Innovación under project / funding scheme PID2020-117191RB-I00 / AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

El propòsit d'aquest projecte és dissenyar i implementar un eina de software sobre un sistema ja existent amb la intenció de millorar l'aprofitament i utilització d'aquest software per part dels estudis de videojocs de la companyia. Aquesta eina pretén automatitzar mitjançant un concepte similar a DevOps Pipelines la generació de proves recurrents i donar a l'usuari del sistema una capa més de personalització en el control d'aquests tests, tot això mantenint i suportant totes les funcionalitats anteriors. Addicionalment, en aquest projecte s'apliquen tots els coneixements que ha adquirit l'autor d'aquest treball sobre l'enginyeria del software en el seu pas per la universitat. En aquest projecte es representa des de la planificació del projecte, els objectius a conseguir i la seva implementació i integració amb el sistema que es vol millorar, passant per fer una recerca i anàlisi dels requisits, els esquemes de dades i arquitectura., The purpose of this project is to design and implement a software tool to an already existing system in order to improve the exploitation and usage of this software by the videogame teams of the company. This tool pretends to automate the generation of tests through a similar concept of DevOps Pipelines and provide the user an extra mask of personalization in the test control as well as keeping and supporting all previous functionalities. Furthermore, all the knowledge that the author has acquired about software engineering during his university years have been applied to this project where it contains its planning, the objectives to achieve, its implementation and integration with the system to improve, research and analysis of the requirements, data schemes and the system architecture.

Unit tests are an essential part of every software project, as they allow to evaluate units of code independently. This final degree thesis aims to enhance unit testing methodology in the context of Ficosa, an international automotive company. This will be done integrating unit test tools within the build system used for the projects, to later automate their execution using Jenkins and finally add unit tests to the continuous integration pipelines., Los tests unitarios son una parte esencial de cualquier proyecto de software, ya que sirven para probar unidades de código de manera independiente. Este proyecto de fin de grado pretende mejorar los métodos de test unitario en el contexto de Ficosa, una empresa internacional del sector de la automoción. Esto se hará a partir de la integración de las herramientas de test unitario en el sistema de compilaciones de los proyectos, para posteriormente automatizar su ejecución me-diante Jenkins y añadir los tests unitarios en las cadenas de integración continua.

El Learning Dashboard és una eina de visualització i monitoratge d'assoliment d'objectius d'aprenentatge per equips d'estudiants que realitzen projectes software. Aquest recull i analitza dades de diverses fonts per calcular i visualitzar mètriques, factors de qualitat i indicadors estratègics. Va ser implementat com a una adaptació del projecte Q-Rapids dashboard. Durant aquesta adaptació, alguns components van deixar de ser funcionals. L'objectiu d'aquest projecte és afegir al Learning Dashboard dues funcionalitats ja existents en el Q-Rapids Dashboard que per el moment no han pogut ser adaptades amb èxit. La primera és un component d'alertes, que envia alertes quan uns valors determinats per a les mètriques, els factors i els indicadors estratègics de seguiment són traspassats. La segona funcionalitat és un component de prediccions, que preveu l'evolució de les diferents mètriques, factors i indicadors estratègics., The Learning Dashboard is a tool for visualizing and monitoring the achievement of learning objectives in student's teams developing software projects. The Learning Dashboard gathers and analyzes information from several sources to calculate and visualize metrics, quality factors and strategic indicators. It was implemented as an adaptation of the Q-Rapids dashboard project. During this adaptation/integration, some components stopped being functional. The objective of this project is to implement in the Learning Dashboard two functionalities already existing in the Q-rapids Dashboard that haven't been adapted successfully. The first one is an alert component, which sends alerts when certain values ¿¿for the tracking metrics, factors and strategic indicators are breached. The second one, a prediction component which predicts the evolution of the different indicators.

Knowledge bases are centralized repositories used for developing knowledge-oriented information systems. They are essential for adaptive, specialized knowledge in dialogue systems, supporting up-to-date domain-specific discussions with users. However, designing large-scale knowledge bases presents multiple challenges in data collection and knowledge exploitation. Knowledge graphs provide various research opportunities by integrating decentralized data and generating advanced knowledge. Our contribution presents a knowledge base in the form of a knowledge graph for extended knowledge generation for mobile apps and features extracted from related natural language documents. Our work encompasses the knowledge graph completion and deductive and inductive knowledge requests. We evaluated the effectiveness and performance of these knowledge strategies, which can be used as on-demand knowledge requests used by third-party software systems., With the support from the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund. This paper has been funded by the Spanish Ministerio de Ciencia e Innovación under project / funding scheme PID2020-117191RB-I00 / AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

Aquest Treball Final de Grau s'ha centrat en expansió d'una eina desenvolupada per la universitat coneguda com Learning Dashboard. Aquest sistema recopila i mostra diverses mètriques relacionades amb projectes de dues assignatures específiques, com per exemple les tasques assignades a cada membre de l'equip o les línies de codi modificades. Malgrat la valuosa informació que ofereix, els alumnes han mostrat una certa reticència a utilitzar aquest portal, potser a causa de la seva falta d'atractiu visual, la seva interfície no tan intuïtiva o el fet que la seva consulta no influeix directament en la seva qualificació. Per a contrarestar aquesta situació i promoure un ús més freqüent de les mètriques disponibles en el Learning Dashboard, es va proposar integrar aquesta informació en una eina d'organització de projectes denominada Taiga, d'ús obligatori en aquestes assignatures. Aquesta iniciativa buscava proporcionar als alumnes i professors un accés més ràpid i eficient a les mètriques, evitant la necessitat de visitar llocs addicionals no essencials per als seus projectes. El primer intent d'implementar aquesta solució va formar part d'un projecte de final de grau previ. Encara que es va crear un sistema funcional, aquest no va complir amb les expectatives en termes d'usabilitat i visualització, però va servir com un valuós punt de partida i va aportar importants lliçons per al desenvolupament futur. Per tant, aquest projecte es va plantejar com una oportunitat per a millorar i perfeccionar aquesta solució inicial. El nostre objectiu principal ha estat abordar les problemàtiques identificades, millorar la interfície d'usuari i desenvolupar un producte final eficient i fàcil d'usar. Aquest producte final té com a objectiu ser utilitzat en les assignatures de PES i ASW, facilitant l'accés a les mètriques dels projectes i optimitzant així el procés d'aprenentatge i avaluació., This Final Degree Project has focused on the expansion of a tool developed by the university known as earning Dashboard. This system compiles and displays various metrics related to projects from two specific subjects, such as the tasks assigned to each team member and the lines of code modified. Despite the valuable information it offers, students have shown some reluctance to use this portal, perhaps due to its lack of visual appeal, its non-intuitive interface, or the fact that its consultation does not directly influence their grade. To counteract this situation and promote more frequent use of the metrics available in Learning Dashboard, it was proposed to integrate this information into a mandatory software projects organization tool called Taiga. This initiative sought to provide students and teachers with quicker and more efficient access to the metrics, avoiding the need to visit additional sites non-essential for their projects. The first attempt to implement this solution was part of a previous final degree project. Although a functional system was created, it did not meet expectations in terms of usability and visualization, but it served as a valuable starting point and provided important lessons for future development. Therefore, this project was posed as an opportunity to improve and perfect this initial solution. Our main objective has been to address the identified problems, improve the user interface, and develop an efficient and easy-to-use final product. This final product aims to be used in the PES and ASW subjects, facilitating access to project metrics and thus optimizing the learning and evaluation process.

Previous research has shown the feasibility of using goal models as a notation to describe data structures. Such results motivated further empirical research to elucidate the extent to which teaching data structures using goal models helps students to better understand (i.e., describe and compare) data structures, and be more efficient when solving their assignments. This paper reports on the experience of introducing goal models for teaching data structures in an undergraduate computer science program during the Autumn semester 2022. We report on the used teaching materials, partial results regarding students’ performance, and our vision for building a platform for teaching data structures with goal models., Peer Reviewed, Postprint (published version)

Software project dashboards provide advanced visualization capabilities and analysis techniques (such as predictions, simulations, what-if analysis, and alerts). They operate by gathering data from several sources (such as code repositories, project management tools, and quality checkers), translating this data to metrics, and next combining metrics to assemble indicators. It becomes crucial to support software project dashboards with a versatile, efficient, and powerful data layer. The purpose of this Final Master Thesis is to design and implement this data layer with a Knowledge Graph data model, thus representing the main concepts managed by the dashboard and the relationships between them. This new data layer shall be integrated into the Q-Rapids (an acronym for Quality-aware Rapid Software Development) dashboard, a particular software project dashboard implemented by the GESSI UPC Research Group, in a non-disruptive manner. The TFM will be evaluated using a customization of the Q-Rapids dashboard in the teaching domain.

El auge de los sistemas de machine learning (ML), la mejora de sus capacidades y el mayor tamaño de los sistemas, ha incrementado el impacto medioambiental de los modelos ML. Sin embargo, la información sobre cómo se mide, comunica y evalúa la huella de carbono de los modelos de ML es escasa. Este proyecto, basado en un análisis de 1.417 modelos de ML y conjuntos de datos asociados en Hugging Face, el repositorio más popular para modelos de ML preentrenados, tiene como objetivo proporcionar una solución integrada para comprender, informar y optimizar la eficiencia de carbono de los modelos de ML. Además, implementamos una aplicación web que genera etiquetas de eficiencia energética para modelos de ML y permite visualizar sus emisiones de carbono. Con menos del 1% de los modelos en Hugging Face proporcionando información sobre las emisiones de carbono, el proyecto subraya la necesidad de mejorar las prácticas de reporte energético y la promoción del desarrollo de modelos eficientes en carbono dentro de la comunidad Hugging Face. Para abordar esta cuestión, ofrecemos una herramienta web que produce etiquetas de eficiencia energética para modelos de ML, una contribución que fomenta la transparencia y el desarrollo de modelos sostenibles dentro de la comunidad de ML. Permite la creación de etiquetas energéticas, al tiempo que proporciona valiosas visualizaciones de los datos de emisiones de carbono. Esta solución integrada constituye un paso importante hacia prácticas de IA más sostenibles medioambientalmente., The rise of machine learning (ML) systems has increased their environmental impact due to the enhanced capabilities and larger model sizes. However, information about how the carbon footprint of ML models is measured, reported, and evaluated remains scarce and scattered. Aims: This project, based on an analysis of 1,417 ML models and associated datasets on Hugging Face, the most popular repository for pretrained ML models, aims to provide an integrated solution for understanding, reporting, and optimizing the carbon efficiency of ML models. Moreover, we implement a web-based application that generates energy efficiency labels for ML models and visualizes their carbon emissions. With less than 1% of models on Hugging Face currently reporting carbon emissions, the project underscores the need for improved energy reporting practices and the promotion of carbon-efficient model development within the Hugging Face community. To address this, we offer a web-based tool that produces energy efficiency labels for ML models, a contribution that encourages transparency and sustainable model development within the ML community. It enables the creation of the energy labels, while also providing valuable visualizations of carbon emissions data. This integrated solution serves as an important step towards more environmentally sustainable AI practices.

Background: When using deep learning models, one of the most critical vulnerabilities is their exposure to adversarial inputs, which can cause wrong decisions (e.g., incorrect classification of an image) with minor perturbations. To address this vulnerability, it becomes necessary to retrain the affected model against adversarial inputs as part of the software testing process. In order to make this process energy efficient, data scientists need support on which are the best guidance metrics for reducing the adversarial inputs to create and use during testing, as well as optimal dataset configurations. Aim: We examined six guidance metrics for retraining deep learning models, specifically with convolutional neural network architecture, and three retraining configurations. Our goal is to improve the convolutional neural networks against the attack of adversarial inputs with regard to the accuracy, resource utilization and execution time from the point of view of a data scientist in the context of image classification. Method: We conducted an empirical study using five datasets for image classification. We explore: (a) the accuracy, resource utilization, and execution time of retraining convolutional neural networks with the guidance of six different guidance metrics (neuron coverage, likelihood-based surprise adequacy, distance-based surprise adequacy, DeepGini, softmax entropy and random), (b) the accuracy and resource utilization of retraining convolutional neural networks with three different configurations (one-step adversarial retraining, adversarial retraining and adversarial fine-tuning). Results: We reveal that adversarial retraining from original model weights, and by ordering with uncertainty metrics, gives the best model w.r.t. accuracy, resource utilization, and execution time. Conclusions: Although more studies are necessary, we recommend data scientists use the above configuration and metrics to deal with the vulnerability to adversarial inputs of deep l, This work was supported by the GAISSA Spanish research project (ref. TED2021-130923B-I00; MCIN/AEI/10.13039/501100011033), the “UNAM-DGECI: Iniciación a la Investigación (verano otoño 2021)” scholarship provided by Universidad Nacional Autónoma de México (UNAM), the “Beatriz Galindo” Spanish Program BEAGAL18/00064, the Austrian Science Fund (FWF): I 4701-N and the project Continuous Testing in Production (ConTest) funded by the Austrian Research Promotion Agency (FFG): 888127., Peer Reviewed, Objectius de Desenvolupament Sostenible::7 - Energia Assequible i No Contaminant, Objectius de Desenvolupament Sostenible::13 - Acció per al Clima, Postprint (published version)

Quality frameworks have been used in requirements engineering (RE) for a long time to help elicit and document quality requirements. However, existing quality frameworks have major issues that hamper their applicability, particularly in RE, but also in other fields such as the design of digital systems. In this paper, we discuss the issues of existing quality frameworks and propose a new quality model, which has been designed for application as a quality framework in RE as well as in the design of digital systems. We present the rationale and requirements for our new model, introduce the model and sketch its application. Our work contributes to the improvement of quality frameworks used in RE and Digital Design., Peer Reviewed, Postprint (published version)

La necessitat de disposar d'entorns d'execució deterministes i alineats amb l'evolució del codi per als entorns de desenvolupament i integració continua de software és una necessitat que cada cop és més urgent a Ficosa. D'aquesta forma, el projecte investigarà i determinarà com es poden integrar una sèrie de tecnologies que ja s'utilitzen a Ficosa amb tot el procés de generació de paquets software. A més, l'objectiu és que el fet de satisfer aquesta necessitat tingui un impacte mínim sobre la forma de treballar de l'empresa i els processos que ja estan establerts i respecten els estàndards arbitrats., The need to have deterministic execution environments aligned with code evolution for software development and continuous integration is an increasingly urgent requirement at Ficosa. In this way, the project will investigate and determine how a series of technologies already used at Ficosa can be integrated into the software package generation process. Additionally, the objective is to ensure that meeting this need has minimal impact on the company's way of working and the processes that are already established and adhere to the based standards.

Chatbots4Mobile is a research project from the GESSI research group (UPC-BarcelonaTech) which aims at designing and developing a task oriented, knowledge based conversational agent to support mobile users in the process of managing and integrating the functionalities exposed by their own application portfolio. To support the design of the required knowledge base, the project focuses on the application of Natural Language Processing (NLP) techniques to infer extended knowledge about the features exposed by a subset of mobile applications, including feature extraction from app-related documents, syntactic and semantic similarity analysis between features, and intent/entity classification focused on functionality identification from user requests. As next steps, we are focusing on the evaluation of embedded linguistic knowledge in large language models, as well as the application of granular sentiment analysis techniques to discern biased documents and process sentiment-based user feedback., With the support from the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia and the European Social Fund. This paper has been funded by the Spanish Ministerio de Ciencia e Innovación under project / funding scheme PID2020-117191RB-I00 / AEI/10.13039/501100011033., Peer Reviewed, Postprint (published version)

When space missions plan scientific actions for robotic spacecraft to execute, they frequently do so within a geometric context called an opportunity. Although there are geometric software libraries that let users write code to search for opportunities, they require knowledge of algorithms and imperative programming languages, which is a condition that might exclude a potentially large population of scientists. Additionally, there might be more user-friendly software systems for scientists to model and search for opportunities, but those might exclude other missions due to export concerns or an inability to maintain such software due to lack of staff or funding. To address these concerns, we designed two different computer languages to model opportunities. In this paper, we present these two languages, our study to evaluate their relative readability and usability, and results obtained in our research along with an interpretation of the same. The metric for this study has been a questionnaire with active exercises, statements with corresponding responses on a Likert scale, and open-ended questions to elicit qualitative responses. The study’s quantitative results provide us with relative and absolute quantification of the usability and readability of each language, while the study’s qualitative results help us direct future language design decisions., Peer Reviewed, Postprint (author's final draft)

The evaluation of Deep Learning (DL) models has traditionally focused on criteria such as accuracy, F1 score, and related measures. The increasing availability of high computational power environments allows the creation of deeper and more complex models. However, the computations needed to train such models entail a large carbon footprint. In this work, we study the relations between DL model architectures and their environmental impact in terms of energy consumed and CO2 emissions produced during training by means of an empirical study using Deep Convolutional Neural Networks. Concretely, we study: (i) the impact of the architecture and the location where the computations are hosted on the energy consumption and emissions produced; (ii) the trade-off between accuracy and energy efficiency; and (iii) the difference on the method of measurement of the energy consumed using software-based and hardware-based tools., This work has been supported by the Spanish project PID2020-117191RB-I00 funded by MCIN/AEI/10.13039/501100011033., Peer Reviewed, Postprint (published version)

As the use of Application Programming Interfaces (APIs) is increasingly growing, their evolution becomes more challenging in terms of the service provided according to consumers' needs. In this paper, we address the role of consumers' needs in WAPIs evolution and introduce a process mining pattern-based method to support providers in WAPIs evolution by analyzing and understanding consumers' behavior, imprinted in WAPI usage logs. We take the position that WAPIs' evolution should be mainly usage-based, i.e., the way consumers use them should be one of the main drivers of their changes. We start by characterizing the structural relationships between endpoints, and next, we summarize these relationships into a set of behavioral patterns (i.e., usage patterns whose occurrences indicate specific consumers' behavior like repetitive or consecutive calls), that can potentially imply the need for changes (e.g., creating new parameters for endpoints, merging endpoints). We analyze the logs and extract several metrics for the endpoints and their relationships, to then detect the patterns. We apply our method in two real-world WAPIs from different domains, education, and health, respectively the WAPI of Barcelona School of Informatics at the Polytechnic University of Catalonia (Facultat d'Informàtica de Barcelona, FIB, UPC), and District Health Information Software 2 (DHIS2) WAPI. The feedback from consumers and providers of these WAPIs proved the effectiveness of the detected patterns and confirmed the promising potential of our approach., This paper has been funded by the Spanish Ministerio de Ciencia e Innovación under project/funding scheme PID2020-117191RB-I00/AEI/10.13039/501100011033., Peer Reviewed, Postprint (published version)

L'aprenentatge automàtic, un tipus particular d'intel·ligència artificial, s'utilitza cada vegada més en els sistemes de software modern. A més de requerir molta energia, el seu ús augmenta les emissions de CO2, té un elevat cost financer i consumeix molts recursos. Encara que existeixen nombrosos estudis que analitzen les possibilitats del ML, molt pocs d'ells se centren en les seves característiques ecològiques, com l'ús d'energia o la seva petjada de carboni. Aquest estudi tracta d'augmentar la conscienciació dels costos associats al desplegament d'un sistema de ML en algunes de les plataformes d'allotjament en el núvol més avançades. Per a això, duem a terme un ampli estudi empíric per a avaluar i contrastar les emissions i la latencia de quatre proveïdors de núvol diferents, Amazon Web Services, Azure, Heroku i Railway, Machine learning, a particular type of artificial intelligence, is increasingly being used in modern software systems. In addition to requiring a lot of energy, its use raises CO2 emissions, has a high financial cost, and consumes many resources. Even while there are numerous studies that look at the possibilities of ML, very few of them concentrate on its green features, like energy consumption or its carbon footprint. This bachelor thesis tries to increase understanding of the costs associated with deploying an ML system into some of the state-of-the-art cloud hosting platforms. In order to do this, we conduct an extensive empirical study to assess and contrast the emissions and latency of four different cloud providers: Amazon Web Services, Azure, Heroku, and Railway.

Disposem d'un Tauler de Comandament Estratègic per al seguiment de projectes software desenvolupats per equips d'estudiants, el Learning Dashboard, amb una interfície web desenvolupada a mida que proporciona una sèrie de visualitzacions sobre diferents mètriques i indicadors de progrés. En aquest projecte es tracta d'avaluar la possibilitat de complementar o substituir aquest component per una eina off-the-shelf específica de construcció i configuració de taulers de comandaments, Grafana., We have a Strategic Dashboard for monitoring software projects developed by student teams, Learning Dashboard, with a custom-developed web interface that provides a series of visualisations on different metrics and progress indicators. This project aims to evaluate the possibility of complementing or replacing this component with a specific off-the-shelf tool for building and configuring dashboards, Grafana.

The main goal of this study is to analyze the impact of the model architecture and training strategy when building green computer vision systems. We divide this goal into two research questions. In the first one, we analyze the impact of the model architecture on achieving greener models while keeping accuracy at optimal levels. In the second one, we study the influence of training strategy on producing greener systems.

Issue trackers, such as Jira, have become the prevalent collaborative tools in software engineering for managing issues, such as requirements, development tasks, and software bugs. However, issue trackers inherently focus on the lifecycle of single issues, although issues have and express dependencies on other issues that constitute issue dependency networks in large complex collaborative projects. The objective of this study is to develop supportive solutions for the improved management of dependent issues in an issue tracker. This study follows the Design Science methodology, consisting of eliciting drawbacks and constructing and evaluating a solution and system. The study was carried out in the context of The Qt Company's Jira, which exemplifies an actively used, almost two-decade-old issue tracker with over 100,000 issues. The drawbacks capture how users operate with issue trackers to handle issue information in large, collaborative, and long-lived projects. The basis of the solution is to keep issues and dependencies as separate objects and automatically construct an issue graph. Dependency detections complement the issue graph by proposing missing dependencies, while consistency checks and diagnoses identify conflicting issue priorities and release assignments. Jira's plugin and service-based system architecture realize the functional and quality concerns of the system implementation. We show how to adopt the intelligent supporting techniques of an issue tracker in a complex use context and a large data-set. The solution considers an integrated and holistic system view, practical applicability and utility, and the practical characteristics of issue data, such as inherent incompleteness., The work presented in this paper has been conducted within the scope of the Horizon 2020 project OpenReq, which is supported by the European Union under Grant Nr. 732463. We are grateful for the provision of the Finnish computing infrastructure to carry out the tests (persistent identifier urn:nbn:fi:research-infras-2016072533). This paper has been funded by the Spanish Ministerio de Ciencia e Innovacionúnder project / funding scheme PID2020-117191RB-I00 / AEI/10.13039/501100011033., Peer Reviewed, Postprint (published version)