Bernardo Leite

Advancing Educational Technology • PhD

Hey 👋! From an early age, I have been passionate about computers and how they operate (after cleaning many viruses!). Today, I am a researcher focused on the use of AI and natural language processing for educational applications. I am currently an AI Engineer at Porto Editora.

I hold a PhD in Informatics Engineering from the Faculty of Engineering - University of Porto (FEUP), where I focused my research on Question Generation.

During this time, I also served as an Invited Assistant at FEUP, teaching Algorithms and Data Structures (1st semester) and Algorithm Design (2nd semester).

I am open to new collaborations. Feel free to reach me at bernardo.leite.lab@gmail.com.

My projects

question-generation-t5-pytorch-lightning: Question Generation with T5 model using Transformers and Pytorch Lightning
gui-qg-llm-pt: Interface for Portuguese Question Generation using LLMs
bernardoleite.github.io: This website made with HTML and CSS
question-generation-portuguese: Automatic Question Generation & Difficulty Control for the Portuguese Language (rule-based)
answer-extraction-and-qg: Tool for Answer Extraction and Question Generation
argument-density: Predicting Argument Density from Text (team project)
ArgMine Platform: Web-based platform for annotating argumentative text (team project)

My (selected) publications

Bernardo Leite, Henrique Lopes Cardoso. Advancing Question Generation with Joint Narrative and Difficulty Control. BEA @ ACL, 2025. slides
Bernardo Leite, Tomás Freitas Osório, Henrique Lopes Cardoso. FairytaleQA Translated: Enabling Educational Question and Answer Generation in Less-Resourced Languages. ECTEL, 2024. slides | code & data
Bernardo Leite, Henrique Lopes Cardoso. On Few-Shot Prompting for Controllable Question-Answer Generation in Narrative Comprehension. CSEDU, 2024. slides | code
Tomás Osório, Bernardo Leite, Henrique Lopes Cardoso, Luís Gomes, João Rodrigues, Rodrigo Santos, António Branco. PORTULAN ExtraGLUE Datasets and Models: Kick-starting a Benchmark for the Neural Processing of Portuguese. BUCC @ LREC-COLING, 2024. data
Bernardo Leite, Henrique Lopes Cardoso. Towards Enriched Controllability for Educational Question Generation. AIED, 2023. slides | code
Bernardo Leite, Henrique Lopes Cardoso. Do Rules Still Rule? Comprehensive Evaluation of a Rule-Based Question Generation System. CSEDU, 2023. slides | code
Bernardo Leite, Henrique Lopes Cardoso. Neural Question Generation for the Portuguese Language: A Preliminary Study. EPIA, 2022. slides | code
(best paper award!) Gil Rocha, Bernardo Leite, Luís Trigo, Henrique Lopes Cardoso, Rui Sousa-Silva, Paula Carvalho, Bruno Martins, Miguel Won. Predicting Argument Density from Multiple Annotations. NLDB, 2022. slides | code
Pedro Azevedo, Bernardo Leite, Henrique Lopes Cardoso, Daniel Castro Silva, Luís Paulo Reis. Exploring NLP and Information Extraction to Jointly Address Question Generation and Answering. AIAI, 2020. slides | code
...full publication list...