Unlearning spurious correlations (TMLR)
Machine learning models often rely on spurious correlations — features that are strongly associated with class labels but lack any causal connection. For exa...
Posts by Category
blog
Knowledge Editing in Large Language Models (NAACL’25)
In our ever-changing world, knowledge of language models (LLMs) needs to be constantly updated. Facts change and LLMs need to reflect this.
LLMs for Generating and Evaluating Counterfactuals (EMNLP’24)
“What minimal changes to this text would cause the text classifier to change its prediction?” Counterfactual texts, i.e. minimal changes to inputs that alter...
InfoLossQA – Characterizing and Recovering Information Loss in Text Simplification (ACL’24)
Text simplification aims to make texts more accessible to a broader audience. Simplification, however, is inherently a lossy process. Additionally, one simpl...
PhD Award Meike Nauta
Meike Nauta wins the Overijssel PhD Award for her thesis “Explainable AI and Interpretable Computer Vision - From Oversight to Insight.” The award is recogni...
Start of the Marflex Project Marcel – Student Chatbot
The MarFlex initiative aims to make study programs and processes more flexible to cater for (more) diverse needs of students. Under this umbrella, we will d...
‘The Queen of England is not England’s Queen.’ Factual coherency of PLMs (EACL’24)
When humans know “Rome is the capital of Italy”, they also know “The capital of Italy is Rome”. This means, if humans know a fact, they can be either queried...
Start of the Pioneers in Healthcare project ReBreCa
Pioneers in Healthcare is an innovation fund awarded by University of Twente, Saxion, MST, ZGT and Deventer Hospital awarding projects with impact on health ...
Paper ‘Give Me the Facts!’ at EMNLP’23
How do we know what Language Models know? And what are obstacles in using them as knowledge bases? In our recent paper presented at EMNLP, we survey methods ...
PIP-Net - interpretable computer vision at CVPR’23
Interpretable methods using prototypical patches help AI explain its reasoning to humans. However, current prototype-based methods may not align with human v...