Author Archive: marisabs

Calouste Gulbenkian Foundation supports 24 outstanding talents from the University of Porto

Posted on

The Calouste Gulbenkian Foundation will support 24 outstanding students from the University of Porto through the Gulbenkian Novos Talentos Scholarships 2025. The program aims to recognize promising young talents, fostering academic and scientific development, and encouraging research and innovation across various fields of knowledge.

Among the awardees, we highlight students from Informatics Engineering and Artificial Intelligence & Data Science, reinforcing the University of Porto’s position as a hub of technological and scientific talent:

António Rodrigues, from the Bachelor in Informatics and Computing Engineering (FEUP/FCUP), recognized for excellence in his academic trajectory.

Pedro Dinis Jorge, from the Bachelor in Artificial Intelligence and Data Science (FCUP/FEUP), also acknowledged for his academic performance and research potential.

The scholarships can reach up to €3,500 per student, covering research expenses, tuition fees, and complementary activities such as participation in conferences, workshops, or international internships. Beyond financial support, the program provides personalized mentoring and networking sessions, promoting the integration of these young talents into the scientific and professional community.

This initiative by the Calouste Gulbenkian Foundation represents a valuable incentive for students to continue developing advanced skills in strategic areas such as computing, AI, and data science, contributing to technological and scientific advancement at both national and international levels.

The University of Porto congratulates the awardees and reaffirms its commitment to supporting and promoting the successes of its students, preparing them to become leaders in their fields.

ICGI’2025 highlights the excellence of DEI researchers

Posted on

This year, the International Conference on Graphics and Interaction 2025 (ICGI’2025) took place on November 13-14, in Sintra, jointly organised by ISTAR_Iscte, Iscte-Sintra and the Eurographics Portuguese Chapter (GPCG).

The conference brought together researchers, teachers and professionals from the fields of Computer Graphics, Image Processing, Computer Vision and Human-Computer Interaction, promoting the dissemination of completed or ongoing work and encouraging the exchange of experiences between the academic and industrial sectors and society.

One of the highlights was the presentation of the article “Materializer: Material Changing Haptics for VR“, authored by the former M.EIC student Henrique Mota Ribeiro, together with Daniel Mendes and Rui Rodrigues, both lecturers at DEI. The work received two distinctions of excellence, winning the Best Paper and Best Presentation awards, reinforcing the scientific and innovative relevance of the project.

The programme also featured prestigious keynotes such as Corina Sas (Lancaster University) and Daniel Gonçalves (IST/UL), leading experts in the areas covered by the event, and an exhibition open to industry, where brands and institutions presented innovative products and services.

DEI Talks | “From Digital Media to Generative AI: The Augmented Environments Lab at Georgia Tech” by Prof. Jay Bolter

Posted on

The talk “From Digital Media to Generative AI: The Augmented Environments Lab at Georgia Tech” will be presented by Prof. Jay Bolter on November the 20th, at 16:30, in room D101 (previously announced in room I-105) and will be moderated by Prof. António Coelho (DEI).

The talk will also be broadcast online:

DEI Talks | “From Digital Media to Generative AI: The Augmented Environments Lab at Georgia Tech” by Prof. Jay Bolter | Meeting-Join | Microsoft Teams

Abstract:

“For more than 25 years, I have been working with students and colleagues at the Georgia Institute of Technology on the history, theory, and practice of digital media. Highlights of that work include projects in our Augmented Environments Lab exploring virtual, augmented, and mixed reality—particularly in the contexts of cultural heritage, entertainment, and personal expression. My interest in the role of digital media within the broader history of media has led to papers and books such as Remediation: Understanding New Media (1999), co-authored with Richard Grusin, and Reality Media: Augmented and Virtual Reality (2021), co-authored by Blair MacIntyre and Maria Engberg. In my presentation, I will review this body of work and suggest how it connects to current research with students, with a particular focus on generative AI. In harvesting billions of words and images from the internet, generative AI performs an algorithmic remix—or remediation—of all the digitized and digital media of the past.”

About the Speaker:

Jay David Bolter is now Professor Emeritus at Georgia Tech. In 2025, he retired as the Wesley Chair of New Media and co-Director of the Augmented Environments Lab. He has lectured extensively in North America and Europe and served as a guest professor at Malmö University in Sweden. His books include Remediation (with Richard Grusin), Windows and Mirrors (with Diane Gromala), The Digital Plenitude, and Reality Media (with Blair MacIntyre and Maria Engberg). CV (PDF)

PhD Defense in Informatics Engineering (ProDEI): ”Educational Question Generation with Narrative and Difficulty Control: A Special Focus on Portuguese”

Posted on

Candidate:
Bernardo José Coelho Leite

Date, Time and Location:
17 November 2025, 14:0, na Sala de Atos da Faculdade de Engenharia da Universidade do Porto

President of the Jury:

Pedro Nuno Ferreira da Rosa da Cruz Diniz (PhD), Full Professor in the Department of Informatics Engineering, Faculdade de Engenharia, Universidade do Porto.

Members:

Hugo Ricardo Gonçalo Oliveira (PhD), Associate Professor in the Department of Informatics Engineering, Faculdade de Ciências e Tecnologia, Universidade de Coimbra;

Maria Luísa Torres Ribeiro Marques da Silva Coheur (PhD), Associate Professor in the Department of Informatics Engineering, Instituto Superior Técnico, Universidade de Lisboa;

Luís Paulo Gonçalves dos Reis (PhD), Associate Professor with Habilitation in the Department of Informatics Engineering, Faculdade de Engenharia, Universidade do Porto;

Henrique Daniel de Avelar Lopes Cardoso (PhD), Associate Professor in the Department of Informatics Engineering, Faculdade de Engenharia, Universidade do Porto (Supervisor).

Abstract:

Humans pose questions all the time, and efforts to create AI systems to do the same have been developed. This task, known as Question Generation (QG), is a subfield of natural language generation that aims to automatically produce relevant and grammatically correct questions from a given input, such as text. A key motivation for QG is to support time-consuming tasks like the manual creation of educational questions by teachers. While QG systems have significantly improved, grammatical accuracy alone does not ensure educational value. Consequently, the adoption of QG tools in educational contexts remains limited.

This thesis is driven by three key challenges in QG: (1) the trustworthiness of AI-generated questions; (2) the limited controllability; (3) restricted applicability in less-resourced languages. To address these challenges, we focus on generating open-ended and multiple-choice reading comprehension questions from narrative texts for elementary school students. For challenge 1, we analyze and report the quality of generated questions, identifying both successful and failed cases. For challenge 2, we enhance controllable generation mechanisms by incorporating multiple attributes, such as narrative elements, explicitness, and difficulty, into the generated questions. Challenge 3 is addressed through a special focus on Portuguese, a morphologically rich language that remains underrepresented in QG research.

Our methodology spans from early rule-based and neural approaches to more advanced controllable QG techniques, including fine-tuning, zero- and few-shot prompting with both small and large language models. This offers a comprehensive view of the evolution and performance of QG systems across different stages. We contribute by systematically applying and adapting current QG techniques. We develop case studies that explore controllability and educational relevance, providing comprehensive analyses of question quality, and releasing new QG models and datasets tailored to less-resourced languages such as Portuguese. Evaluation combines automatic metrics with human-centered assessments involving experts, teachers, and students, whose input provides critical insights into the usefulness and effectiveness of the generated questions.

The results show that it is possible to generate well-formulated and answerable questions with controllable attributes. Although machine-generated questions approach the quality of humanauthored ones, semantic issues still arise. In addition, generating MCQs with answer options that are effective for students remains a challenge. These findings highlight the ongoing need for research in educational QG, especially in supporting less-resourced languages and enhancing the reliability of automated generation systems.

SINF 2025: Another inspiring edition of Semana de Informática wraps up

Posted on

By Álvaro Paralta, SICC, FEUP

“Lectures, workshops and problem-solving sessions. These were some of the activities at the latest edition of Semana de Informática (SINF), organised by the Informatics Group of the Student Association of the Faculty of Engineering of the University of Porto (NIAEFEUP). The initiative took place between 21 and 24 October at the Faculty of Engineering and was attended by more than 20 companies.

Over the four days, the companies interacted freely with FEUP students from their respective stands, distributed along the main corridor of the faculty. But that’s not all. In addition to the job fair itself, the programme also included workshops, lectures, problem-solving sessions and a networking dinner between the companies with the highest level of sponsorship at the event and the invited students.

According to Diogo Fernandes, president of NIAEFEUP, the overall assessment of the event is “very positive”. “We got the impression that the companies, in general, were very satisfied. And the feedback we received, both from trainers and speakers, was that the students participated actively in the activities – they were always very interested and engaged,” he comments.
“For us, organising a SINF means having the opportunity to bring new topics to students, as well as providing them with another opportunity to interact with companies. This is clearly something in which FEUP excels, in general – the ease with which we are able to reach the business community. There are always many job fairs and activities with companies that enhance this relationship. And having an activity of this type dedicated solely to Computer Engineering is very important,” reflects Diogo Fernandes.

Among the topics covered at SINF were Artificial Intelligence, Digital Transformation, Cybersecurity, Quantum Computing, Career Development and Entrepreneurship. “We have noticed that this has been an area that has been emerging more and more – not only because of growing demand from students, but also because of what we have seen at the events we have participated in,” explains Diogo.

There was also room for new features, such as problem-solving sessions and a networking dinner. “The contact that a company has with a student through a job fair is usually relatively quick and does not allow for much discussion or in-depth conversation. With this in mind, we prepared these activities, which we believe allowed for more detailed contact between the company and the student,” shares the president of NIAEFEUP. “With the problem-solving sessions, we gave companies a chance to closely observe the students’ ability to solve challenges. On the one hand, we managed to give students a more relaxed moment; on the other hand, companies were able to see their skills in a very practical way – not only in terms of solving the challenge itself, but also their teamwork, communication and leadership skills,” explains Diogo Fernandes.

The networking dinner was also developed in a more personalised way. The companies with the highest level of sponsorship were able to directly invite students who had positively impressed them during the interaction at their respective stands. This allowed the relationship between the employer and the potential candidate to be deepened, enabling mutual interests and opportunities to be explored in a more informal way and with more time.

NIAEFEUP organised this edition of SINF in the aftermath of the National Meeting of Computer Science Students, which took place in April at FEUP. “This was an added challenge because it was a very large event to which we were all very dedicated. This meant that we ended up starting SINF a little later,” shares the student. “It was challenging, but we already had a well-defined plan. We ended up gaining a lot of experience from the meeting, and that helped us.”

With their eyes already set on the future, NIAEFEUP’s goals include organising more training sessions with companies throughout the year. “We want to be more involved in the student training process, meeting their needs. And that’s what we’re working on,” concludes Diogo.

More information here.”

ECMLPKDD 2025 – European Conference on Machine Learning and Knowledge Discovery in Databases

Posted on

The European Conference on Machine Learning and Knowledge Discovery in Databases (ECMLPKDD), the largest European event dedicated to Machine Learning and one of the most important worldwide in the field of Artificial Intelligence, took place in Porto between 15 and 19 September 2025.

The organisation counted on the active participation of several students and researchers from the Department of Informatics Engineering (DEI) of the Faculty of Engineering of the University of Porto (FEUP), notably Carlos Soares (General Co-Chair) and João Mendes Moreira (Workshops Co-Chair).

Held at Alfândega do Porto, the conference brought together more than 1,300 participants from 60 countries, including around 450 students. The programme featured seven keynote speeches by some of the world’s leading researchers in the field, including Pedro Domingos (University of Washington), Cynthia Rudin (Duke University), Mirella Lapata (University of Edinburgh), Francisco Herrera (University of Granada), Sašo Džeroski (Jožef Stefan Institute) and Nuria Oliver (ELLIS Alicante – Institute of Humanity-Centric AI).

The scientific programme included around 400 accepted papers, in addition to numerous presentations at the 32 workshops associated with the event, reflecting the vitality and diversity of the European Machine Learning and Data Science community.

ECMLPKDD 2025 was supported by several national and international companies, including BNP Paribas, EDF, Google, ASML, NOS, NEC, Amazon, AstraZeneca and Banco de Portugal, as well as an institutional partnership with Porto Digital.

More information on the ECMLPKDD website.

DEI Talks | “smtgcc: Using an SMT solver to find bugs in GCC” by Krister Walfridsson

Posted on

The talk “smtgcc: Using an SMT solver to find bugs in GCC” will be presented by Krister Walfridsson on December the 4th, at 16:00, online:

Join the meeting 
Meeting ID: 373 912 942 228 7
Passcode: XS9M8dT3

Abstract:

“SMT solvers are increasingly effective for finding compiler bugs and validating optimizations. This talk presents smtgcc, a translation-validation tool for GCC. It is similar to Alive2 for LLVM, but smtgcc’s approach diverges from Alive2 because GCC and LLVM follow different design choices. I will explain how smtgcc works and discuss issues in formalizing the semantics of GIMPLE, GCC’s IR.”

About the Speaker:

Krister Walfridsson became involved with the GCC project while studying at university in the mid-1990s. Since completing his studies, he has worked with both compilers and operating systems in various embedded environments. Most recently, he spent 10 years at Arm as a principal compiler engineer in the Mali GPU team. He is currently taking a few years off to work on personal projects and to dance.

DEI Talks | “Knowledge Graphs + AI: The Evolution of Automated GitHub Issue Resolution” by Prof. He Ye (University College London)

Posted on

The talk entitled “Knowledge Graphs + AI: The Evolution of Automated GitHub Issue Resolution“, will be presented by Prof. He Ye on November 17th, at 14:30, in room B008, moderated by Prof. Alexandra Mendes (DEI).

Abstract:

“AI coding agents are becoming increasingly capable, achieving strong results on benchmarks such as SWE-bench. However, most still struggle with real-world challenges such as issue reproduction, precise context retrieval from large codebases, and the high cost of LLMs. In this talk, I will introduce our recent code agent, Prometheus — a knowledge graph-powered, multi-agent system designed to tackle GitHub issues in practice. Prometheus transforms entire repositories into a unified knowledge graph stored in Neo4j for scalable and structured reasoning. This enables precise, cross-language context retrieval, allowing large language models to generate accurate and efficient fixes. Prometheus delivers robust performance, resolving diverse issues across seven programming languages. I will show how combining LLMs with knowledge graphs can advance automated issue resolution beyond today’s benchmark-driven limits. We have recently transitioned this research into an off-the-shelf product that helps industry resolve software issues automatically.”

About the Speaker:

He Ye is an Assistant Professor at University College London. She previously worked as a Postdoctoral Researcher at Carnegie Mellon University and received her PhD from KTH Royal Institute of Technology. Her research centers on developing the next generation of code agents to automate software engineering tasks, with a focus on codebase context retrieval, automated issue resolution, and code agent memory construction. Beyond academia, she is the co-founder of EuniAI, a startup committed to turning research into real-world solutions that help developers address practical software challenges.

FCUP and FEUP dominated MIUP’s podium

Posted on

By Renata Silva (FCUP) and Mafalda Leite (SICC/FEUP)

” Several teams of students from the Faculty of Sciences (FCUP) and Faculty of Engineering (FEUP) of the University of Porto, shone at the most recent edition of the Inter-University Programming Marathon (MIUP), which took place on 18 October at the University of Aveiro’s Águeda School of Technology and Management. U.Porto dominated the competition podium with gold, silver and bronze medals.

Leading the way were the “Tiny Silly Problem” team, who managed to solve eight of the 10 problems proposed. The team is made up of students Marco Vilas Boas, from the Master’s in Informatics and Computing Engineering (FEUP), Félix Martins, from the Master’s in Artificial Intelligence (FEUP and FCUP), and Patrick Daniel, from the Master’s in Computer Science (FCUP).

One of the competition’s silver medals went to “Long long main“, by students Luís Barbosa, Luís Gonçalves and Sofia Sousa (all from the 3rd year of the Bachelor in Informatics and Computing Engineering at FEUP and FCUP), for managing to solve four of the proposed problems.

In 5th place in the competition were “oneL“, who won bronze, also by solving four of the competition’s challenges. This team includes students Filipe Zheng, from the 2nd year of the Bachelor in Artificial Intelligence and Data Science (FCUP and FEUP), Miguel Rocha and Rayner Sulyak, from the 1st and 2nd year of the Bachelor in Computer Science (FCUP), respectively.

Also taking part in the competition were “The CodeFathers“, students Joana Louro and Luís Santos, from the 3rd year of the Bachelor in Informatics and Computing Engineering, and Luís Santos and Pedro Machado, from the Bachelor in Computer Science, who finished in 13th place.

The teams were coached by Pedro Ribeiro, a lecturer at FCUP’s DCC, and André Restivo, a lecturer at FEUP’s Department of Informatics Engineering (DEI). This is the third year that the FEUP and FCUP teams have competed under the unified banner of the University of Porto. “These results highlight the talent and dedication of our students, as well as the collaboration between FEUP and FCUP. In particular, the winning team, made up of elements from both faculties and from a joint course, shows how co-operation within the University of Porto can lead to excellent results,” says André Restivo (DEI).

Pedro Ribeiro, a lecturer at FCUP, also highlights “the gratifying joint effort of two faculties who realise that UP as a whole is stronger and more united in an area of great prestige”. “Participating in the competitions and all the preparation and training involved is of great importance to the scientific growth of our participants, and we are proud to see a wide range of academic and professional excellence in the current and former competitors,” he adds.

The three medal-winning teams will now represent U.Porto at the Southwestern Europe Regional Contest (SWERC) 2025, which will take place from 21 to 23 November at the Instituto Superior Técnico in Lisbon.

About MIUP

The Inter-University Programming Marathon (MIUP) is a programming competition aimed at higher education students. The competition is run by teams of up to three people, where for five hours the teams will try to solve between nine and 11 different problems, using the programming languages C, C++, Java or Python.
The competition allows participants to test their problem-solving skills, while at the same time enabling students and teachers from Portuguese higher education institutions to socialise and exchange experiences.”

DEI Talks | “Energy-awareness in compute acceleration: The role of FPGAs” by Prof. Shreejith Shanker

Posted on

The talk entitled “Energy-awareness in compute acceleration: The role of FPGAs“, will be presented by Prof. Shreejith Shanker on October 30, at 11:30, in room B012, and will be moderated by Prof. Tiago Carvalho (DEI).

Abstract:

“The talk will cover a set of projects that my team at TCD is working on, spanning embedded and distributed systems to high-performance media workflows, and how FPGAs are enabling an energy-performance trade-off in these applications.”

About the Speaker:

Dr. Shreejith Shanker is an Assistant Professor of Reconfigurable Computing at Trinity College Dublin, Ireland and leads the research group on reconfigurable architectures, accelerators and workflows. His research interests include reconfigurable and adaptive computing architectures, in-network computing, post-production media workflows, design automation tools and distributed embedded systems, with a focus on performance-energy trade-off and hardware-software codesign approaches.