Waarom The Book of Why belangrijk is
Wie denkt dat met artificiële intelligentie alles mogelijk wordt, zolang we maar genoeg data verzamelen, vergist zich. Puur observationele data zeggen niets over het proces waarmee ze gegenereerd worden of over de oorzaken die daarachter schuilgaan.
Automated Reasoning and Learning for Automated Payroll Management
This post is based on the following publication:
Automated Reasoning and Learning for Automated Payroll Management, S. Dumancic, W. Meert, S. Goethals, T. Stuyckens, J. Huygen, K. Denies, IAAI 2021 (IAAI-21 Deployed Application Award).
Innovation Lab
This post is based on the Innovation Lab project. KU Leuven faculty of Engineering’s Innovation Lab is an initiative to enthuse high school students to become engineers and scientists by having them build an actual real-life device.
De Machine Learning Revolutie in de Artificiële Intelligentie
Deze samenvatting is gebaseerd op:
Luc De Raedt, Over Machines die Leren, in: Visie en Vooruitgang. Lessen voor de XXIste eeuw, Universitaire Pers Leuven, 2009. https://people.cs.kuleuven.be/~luc.deraedt/Visie.pdf
Luc De Raedt, De artificiële-intelligentierevolutie en de impact ervan, in: Bart Raymaekers, Wetenschap in een veranderende wereld.
Pitfalls in training and evaluating expected goals (xG) models
Expected goals (xG) measures the quality of a shot attempt in soccer based on several variables. They are among the most used and best understood advanced metric in the world of football, but training an accurate xG model is not as simple as it may seem.
RobBERT: a Dutch Language Model
We have developed RobBERT, a language model pre-trained on Dutch texts. This post will give a brief overview of DTAI’s work in this field.
Valuing actions in soccer
The VAEP framework is a novel data-driven framework for assessing the impact of the individual actions performed by soccer players during a game, which is a crucial aspect of the player recruitment process. Here we group a number of blogposts on the methodology and applications of VAEP.
DeepProbLog
This is a brief overview of DeepProbLog, a neuro-symbolic framework that integrates the probabilistic logic programming language ProbLog with neural networks.
Time Series Clustering
This post is based on the following publications:
COBRAS-TS: A new approach to Semi-Supervised Clustering of Time Series. Van Craenendonck, T., Meert, W., Dumancic, S. & Blockeel, H. Discovery Science, 2018.
Anomaly Detection with User Feedback
An overview of DTAI research on how to handle user feedback for contextual anomaly detection.
