Academic
Table of Contents
Position
I am a post-doctoral researcher working on creative artificial intelligence under supervision of prof. dr. Luc De Raedt at the DTAI research group of the KU Leuven computer science department and the Leuven.AI institute.
Publications
During my academic career, I have published the following 22 papers. A more up-to-date list is available on my Google Scholar profile.
Papers marked in blue are recommended if you would like to gain quick insights into my main research interests.
2024
2023
2022
- 2022RobBERT-2022: Updating a Dutch Language Model to Account for Evolving Language UsePieter Delobelle, Thomas Winters, Bettina BerendtPre-print.
- 2022DeepStochLog: Neural Stochastic Logic ProgrammingThomas Winters, Giuseppe Marra, Robin Manhaeve, Luc De RaedtProceedings of the AAAI Conference on Artificial Intelligence.
2021
- 2021RobBERTje: a Distilled Dutch BERT ModelPieter Delobelle, Thomas Winters, Bettina BerendtComputational Linguistics in the Netherlands Journal.
- 2021Survival of the Wittiest: Evolving Satire with Language ModelsThomas Winters, Pieter DelobelleProceedings of the Twelfth International Conference on Computational Creativity.
- 2021Learning to Rank Generated PortmanteausLara Pollet, Thomas Winters, Pieter DelobelleProceedings of the Twelfth International Conference on Computational Creativity.
- 2021Shape inference and grammar induction for example-based procedural generationGillis Hermans, Thomas Winters, Luc De RaedtProceedings of the Twelfth International Conference on Computational Creativity.
- 2021SandSlide: Automatic Slideshow NormalizationSieben Bocklandt, Gust Verbruggen, Thomas WintersProceedings of the 2021 16th International Conference on Document Analysis and Recognition (ICDAR).
- 2021Computers Learning Humor Is No JokeThomas WintersHarvard Data Science Review.
2020
- 2020Dutch Humor Detection by Generating Negative ExamplesThomas Winters, Pieter DelobelleProceedings of the 32st Benelux Conference on Artificial Intelligence (BNAIC 2020) and the 29th Belgian Dutch Conference on Machine Learning (Benelearn 2020).
- 2020Discovering Textual Structures: Generative Grammar Induction using Template TreesThomas Winters, Luc De RaedtProceedings of the 11th International Conference on Computational Creativity.
- 2020Rosetta Code: Improv in Any LanguagePiotr Mirowski, Kory Mathewson, Boyd Branch, Thomas Winters, Ben Verhoeven, Jenny ElfvingProceedings of the 11th International Conference on Computational Creativity.
- 2020RobBERT: a Dutch RoBERTa-based Language ModelPieter Delobelle, Thomas Winters, Bettina BerendtFindings of EMNLP2020.
2019
- 2019Modelling Mutually Interactive Fictional Character Conversational AgentsThomas WintersProceedings of the 31st Benelux Conference on Artificial Intelligence (BNAIC 2019) and the 28th Belgian Dutch Conference on Machine Learning (Benelearn 2019).
- 2019Playable Experiences at the 15th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE’19)Rook Liu, Claire Christopher, Chris Martens, Santiago Ontañòn, Piotr Mirowski, Kory Mathewson, Thomas Winters, Shaun FarrugiaProceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainments.
- 2019Generating Philosophical Statements using Interpolated Markov Models and Dynamic TemplatesThomas WintersEuropean Summer School in Logic, Language and Information.
- 2019Towards a General Framework for Humor Generation from Rated ExamplesThomas Winters, Vincent Nys, Daniel De SchreyeProceedings of the 10th International Conference on Computational Creativity.
- 2019Automatically Generating Engaging Presentation Slide DecksThomas Winters, Kory MathewsonComputational Intelligence in Music, Sound, Art and Design - 8th International Conference, EvoMUSART.
- 2019Generating Dutch Punning Riddles about Current AffairsThomas Winters29th Meeting of Computational Linguistics in the Netherlands (CLIN 2019).
Education
- Post-doctoral researcher in Creative Artificial IntelligenceKU LeuvenAt the DTAI research group under supervision of prof. dr. Luc De Raedt.2023-present
- PhD in Creative Artificial IntelligenceKU LeuvenAt the DTAI research group under supervision of prof. dr. Luc De Raedt. Thesis topic: "Neuro-Symbolic Creative Artificial Intelligence for Humor".2018-2023
- Advanced Master in Artificial IntelligenceKU LeuvenMagna cum laude.2017-2018
- Master of Engineering in Computer ScienceKU LeuvenSumma cum laude. Top 5% of class.2015-2017
- Erasmus (during Master of Engineering in Computer Science)University of BristolFirst class mark on every unit.2015-2016
- Bachelor of Science in InformaticsKU LeuvenMagna cum laude. First of class.2012-2015
Teaching
During my PhD, I have been a teaching assistant for the following courses:
- Generative AI module @ Applied AI: Academic PerspectivesB-KUL-B30773This module is a 12-hour lecture series about generative AI, which I designed from scratch. It covers text generation, image generation, audio generation as well as discussions on the limits and risks of generative AI.2024-Present
- Artificiële Intelligentie voor iedereenB-KUL-H02D2AAn optional course about AI, available to all bachelors at KU Leuven. I interviewed professors about their AI-related research to create videos as follow-up discussions after their lecture.2021-2022
- Artificiële IntelligentieB-KUL-H06U1AThe main AI course for engineering and informatics bachelor students. I help integrate new material into the exercise sessions, regularly create new exercises with extensive solution slideshows and with videos, and teach all thirteen exercise sessions to multiple groups every year.2018-Present
- Capita Selecta Computer Science: Artificial IntelligenceB-KUL-H05N0AMaster's course where AI students work together on their own project. I was the teaching assistant responsible for the Computational Creativity track of the Contemporary Topics of AI submodule. For this, I proposed the computational creativity topics for student pairs to explore, and coached them in their research.2018-2020
Supervision
I supervised the following students' master's theses:
- Humor Theory-Guided Humor Detection and Comprehension with Machine Learning for Steering Autoregressive Language Models Towards Humor Generation(NL) Humortheorie-geleide humordetectie en humorbegrip met machine learning voor de sturing van autoregressieve taalmodellen richting humorgeneratieVictor De Marez2022-2023
- Asking the Right Question: Generating Difficulty-Ranked Questions from ExamplesJinfu Chen2021-2022
- Comic Art Generation using GANsMarnix Verduyn2021-2022
- Satire Generation with the T5 Model(NL) Satire generatie met het T5 modelMaarten Van den hof2021-2022
- BarLocaliser: Neural-Symbolic Detection of Bars in Bar Charts(NL) BarLocaliser: neuraal-symbolische detectie van staven in staafdiagrammenSimon Kupers2020-2022
- SAGE: a Heuristic-based Acronym Generator(NL) SAGE: een heuristiek-gebaseerde acroniemgeneratorTibo Masselis2020-2021
- Morphological Tokenizers for Dutch Language ModelsNicholas Wellens2020-2021
- Normalization of Slideshows in the Conversion from PDF to PowerPoint(NL) Normalisatie van diavoorstellingen bij de conversie van pdf naar PowerPointSieben Bocklandt2020-2021
- Automatic Ranking Generated Portmanteaus(NL) Automatisch rangschikken van gegenereerde porte-manteausLara Pollet2020-2021
- Example-based Procedural Generation: Shape Inference and Grammar Induction from Voxel StructuresGillis Hermans2019-2020
- Learning User Preference in a Generative Anagram Model(NL) Gebruikersvoorkeuren leren in een generatief anagrammodelSimon Roelandt2019-2020
Awards
- Best Video AwardBNAIC/BENELEARN2020Best video presentation of the Belgian AI conference2020-11-20
- FWO FellowshipFonds Wetenschappelijk OnderzoekPersonal PhD funding for four years of fundamental research2019-10-10