avatar

Cedric Waterschoot

Postdoctoral Researcher
Department of Advanced Computing Sciences, Maastricht University
cedric.waterschoot (at) maastrichtuniversity.nl


About Me

I am a postdoctoral researcher at the Department of Advances Computing Sciences (DACS) at Maastricht University. Previously, I was a PhD candidate at KNAW Meertens Instituut in Amsterdam and Utrecht University. My research interests include Group Recommender Systems, Natural Language Processing (NLP), group modelling and automatic summarization.

I wrote my dissertation on hybrid content moderation of discussions hosted by online news outlets (comment sections). You can read my thesis here. I created computational tools to support the content moderator in their tasks. Specifically, I looked at identifying ‘good’ or ‘constructive’ comments, which news outlets manually highlight to promote better discussions. You can read about my research here (Dutch). Furthermore, I studied how content moderators are currently working alongside AI-based tools as well as the opportunities for future collaboration. My PhD project was part of the larger Better-MODS project.

In the past I have looked at how you can (computationally) identify large-scale trends in online discussions and measure impact/agency of user interaction on news platforms. I am currently still working on automatic summarization of online discussions and using language models to create explanations of (group) recommender systems.

Current Projects

Wie is de trol?

‘Wie is de trol?’ is a game for school students and libraries to learn about online discussions, content moderation and the use of Artificial Intelligence (AI). It is an initiative by researchers from KNAW Meertens Instituut, Tilburg University, together with NEMO Kennislink. Visit the website for a full project description.

The game is being developed by Basten Stokhuyzen, Mathilde Jansen and Cedric Waterschoot. This project is financed by a Dutch Research Agenda ‘Science Communication’ grant (NWO).

Previous Projects

Better-MODS: Better Informing Citizens About Current Debates

The BETTER-MODS project is a collaboration between Tilburg University, the KNAW Meertens Institute and NU.nl. This project aims to make online discussions more accessible and to understand how human moderators work alongside AI in the context of hybrid content moderation. Furthermore, we study if people engage in discussions on news articles and the motivation behind their participation.

Recent work

  1. The Impact of Featuring Comments in Online Discussions
    Cedric Waterschoot, Ernst van den Hemel, Antal van den Bosch
    International Conference on Advances in Social Networks Analysis and Mining (ASONAM2024)
    PDF

  2. The Constructive Conundrum: Computational Approaches to Facilitate Constructive Commenting on Online News Platforms
    Cedric Waterschoot
    PDF PhD Thesis

  3. A Time-Robust Group Recommender for Featured Comments on News Platforms
    Cedric Waterschoot, Antal van den Bosch
    Frontiers in Big Data
    PDF Code

  4. AI tilt discussie naar hoger niveau
    NEMO Kennislink, Mathilde Jansen
    PDF Interview

  5. Agency and dynamics in large datasets: Dutch Islam discourse on social media
    Cedric Waterschoot
    Religion in Motion – Between Borders and Belonging (NGG 2023)
    Presentation

  6. Hybrid moderation in the newsroom: Recommending featured posts to content moderators
    Cedric Waterschoot, Antal van den Bosch
    PDF Preprint

  7. Detecting Minority Arguments for Mutual Understanding: A Moderation Tool for the Online Climate Change Debate
    Cedric Waterschoot, Antal van den Bosch, Ernst van den Hemel
    29th International Conference on Computational Linguistics (COLING 2022)
    PDF Code

  8. Constructive Behavior? No Idea, but we recognize it when we see it
    Cedric Waterschoot, Ernst van den Hemel
    Digital Society Conference 2022
    Presentation

  9. Calculating Argument Diversity in Online Discussion Threads
    Cedric Waterschoot, Antal van den Bosch, Ernst van den Hemel
    3rd Conference on Language, Data and Knowledge (LDK2021)
    PDF Code


Powered by Jekyll and Minimal Light theme.