Gilles Vandewiele graduated magna cum laude as a computer science engineer in June 2016 at Ghent University. Later that year, in August, he joined the Internet and Data Science Lab (IDLab) research group in the Department of Information Technology (INTEC). In February 2021, Gilles obtained his PhD in Computer Science Engineering. Gilles conducts research in the domain of white-box machine learning for critical domains and (semantic) knowledge models. Other research interests are bio-inspired algorithms and sport-related data science in general.
If you cannot find Gilles sitting behind his desk, he is most likely doing a code golf in Python, competing in a data science competition or out playing football with friends or colleagues.
dr. ir. Gilles Vandewiele
gilles.vandewiele (at) ugent (dot) be
PhD Student
Internet and Data Science Lab (IDLab)
Department of Information Technology (INTEC)
University Ghent
Belgium
CV
Publications
- Appearing soon MINDWALC: Mining Interpretable, Discriminative Walks for Classification of Nodes in a Knowledge Graph
Gilles Vandewiele, Bram Steenwinckel, Filip De Turck, Femke Ongenae
BMC Medical Informatics and Decision Making - A decision support system to follow up and diagnose primary headache patients using semantically enriched data
Gilles Vandewiele, Femke De Backere, Kiani Lannoye, Maarten Vanden Berghe, Olivier Janssens, Sofie Van Hoecke, Vincent Keereman, Koen Paemeleire, Femke Ongenae and Filip De Turck
BMC Medical Informatics and Decision Making. - Tslearn, A Machine Learning Toolkit for Time Series Data
Romain Tavenard, Johann Faouzi, Gilles Vandewiele, Felix Divo, Guillaume Androz, Chester Holtz, Marie Payne, Roman Yurchak, Marc Rußwurm, Kushal Kolar, Eli Woods
Journal of Machine Learning Research (Machine Learning Open Source Software) - Training progression in recreational cyclists: no linear dose-response relationship with training load
Kobe Vermeire, Gilles Vandewiele, Kevin Caen, Maarten Lievens, Jan Bourgois, Jan Boone
The Journal of Strength & Conditioning Research - Data Mining in the Development of Mobile Health Apps: Assessing In-App Navigation Through Markov Chain Analysis
Jeroen Stragier, Gilles Vandewiele, Paulien Coppens, Femke Ongenae, Wendy Van den Broeck, Filip De Turck, Lieven De Marez
Journal of Medical Internet Research (JMIR) - Modelling the Prediction of the Session Rate of Perceived Exertion in Soccer: Unravelling the Puzzle of Predictive Indicators
Youri Geurkink, Gilles Vandewiele, Maarten Lievens, Filip de Turck, Femke Ongenae, Stijn PJ Matthys, Jan Boone, and Jan G. Bourgois
International Journal of Sports Physiology and Performance (IJSPP)
- Overly Optimistic Prediction Results on Imbalanced Data: Flaws and Benefits of Applying Over-sampling
Gilles Vandewiele , Isabelle Dehaene, György Kovács, Lucas Sterckx, Olivier Janssens, Femke Ongenae, Femke De Backere, Filip De Turck, Kristien Roelens, Johan Decruyenaere, Sofie Van Hoecke, Thomas Demeester - GENDIS: GENetic DIscovery of Shapelets
Gilles Vandewiele, Femke Ongenae, Filip De Turck
- Workshop Inducing a Decision Tree with Discriminative Paths to Classify Entities in a Knowledge Graph
Gilles Vandewiele, Bram Steenwinckel, Femke Ongenae, Filip De Turck
Presented at Semantics-Powered Data Mining and Analytics (SEPDA)
ISWC 2019 (Auckland, New Zealand) - Workshop CSV2KG: transforming tabular data into semantic knowledge
Bram Steenwinckel, Gilles Vandewiele, Femke Ongenae, Filip De Turck
Presented at Semantic Web Challenge on Tabular Data to Knowledge Graph Matching (SemTab)
ISWC 2019 (Auckland, New Zealand) - Main Track A Critical Look at Studies Applying Over-Sampling on the TPEHGDB Dataset
Gilles Vandewiele, Isabelle Dehaene, Olivier Janssens, Femke Ongenae, Femke De Backere, Filip De Turck, Kristien Roelens, Sofie Van Hoecke, Thomas Demeester
Presented at main track
AIME 2019 (Poznan, Poland)
Nomination for Best Student Paper Award - Main Track Time-to-Birth Prediction Models and the Influence of Expert Opinions
Gilles Vandewiele, Isabelle Dehaene, Olivier Janssens, Femke Ongenae, Femke De Backere, Filip De Turck, Kristien Roelens, Sofie Van Hoecke, Thomas Demeester
Presented at main track
AIME 2019 (Poznan, Poland)
- Main Track Assessing navigation in eHealth apps: a Markov chain analysis of the Start2Cycle app
Jeroen Stragier, Gilles Vandewiele, Paulien Coppens, Lieven De Marez, Wendy Van den Broeck, Filip De Turck, and Femke Ongenae
Presented at main track
ISBNPA 2018 (Hong Kong)
Best presentation award - Main Track Training Impulses And The Relation With Performance Improvement: Not That Straightforward
Kobe Vermeire, Gilles Vandewiele, Jan Bourgois, and Jan Boone
Medicine & Science in Sports & Exercise 50 - Main Track Towards an automated workout compliance model
Jelle De Bock, Gilles Vandewiele, Steven Verstockt, Femke Ongenae, and Filip De Turck
Presented at main track
ISPAS 2018 (Zagreb, Croatia) - Workshop Predicting Train Occupancies based on Query Logs and External Data Sources
Gilles Vandewiele, Pieter Colpaert, Olivier Janssens, Joachim Van Herwegen, Ruben Verborgh, Erik Mannens, Femke Ongenae, and Filip De Turck
Presented at the 7th International Workshop on Location and the Web (LocWeb)
WWW 2017 (Perth, Australia) - Main track Real-time Data Dissemination and Analytics Platform for Challenging IoT Environments.
Glenn Daneels, Esteban Municio, Kathleen Spaey, Gilles Vandewiele, Alexander Dejonghe, Femke Ongenae, Steven Latre, and Jeroen Famaey
Presented at the main track
GIIS 2017 (Reunion Island, France) - Workshop Enabling Training Personalization by Predicting the Session Rate of Perceived Exertion (sRPE)
Gilles Vandewiele, Youri Geurkink, Maarten Lievens, Femke Ongenae, Filip De Turck and Jan Boone
Presented at the Machine Learning for Sport Analytics
ECML 2017 (Skopje, Macedonia) - Main track Enhancing White-Box Machine Learning Processes by Incorporating Semantic Background Knowledge
Gilles Vandewiele
Presented at the PhD Symposium
ESWC 2017 (Portoroz, Slovenia) - Workshop A genetic algorithm for interpretable model extraction from decision tree ensembles
Gilles Vandewiele, Kiani Lannoye, Olivier Janssens, Femke Ongenae, Filip De Turck
Presented at the BDM workshop
PAKDD 2017 (Jeju, South-Korea) - Workshop GENESIM: genetic extraction of a single, interpretable model.
Gilles Vandewiele, Olivier Janssens, Femke Ongenae, Filip De Turck, and Sofie Van Hoecke
Presented at the Interpretable ML for Complex Systems workshop
NIPS 2016 (Barcelona, Spain)
- Identifying the Number of Open Ion Channels with Hidden Markov Models
A write-up on how we almost won the “University of Liverpool — Ion Switching” Kaggle competition. - Explainable, efficient and accurate node classification in Knowledge Graphs
Mining discriminative walks with MINDWALC: Mining Interpretable, Discriminative Walks for Classification of Nodes in a Knowledge Graph. - Google Hash Code 2020: a greedy approach
How we obtained ranking 16/204 in Belgium and 531/10724 globally. - Enriching shapelets with positional information for timeseries classification
How a neat simple trick can boost both predictive performance and interpretability. - Helping Santa plan with Mixed Integer Programming (MIP)
A write-up on how we obtained a silver medal in the “Santa’s Workshop Tour 2019” competition organized by Kaggle. - A simple, but strong AI bot for Tron/Lightriders on a hexagonal grid
A blog post elaborating my solution for the second annual AI competition. - Creating a platform for AI competitions in Python
A small blog post about the most significant aspects of the development of a platform to allow students to submit agents to compete against eachother. - Creating the (nearly) perfect connect-four bot with limited move time and file size
Details of my solution for an AI video game bot competition, where you haJ Booned to write a bot that plays the game Connect Four. - Semantically annotating eHealth data using the SNOMED-CT ontology
It took me quite some time to get a local copy of the SNOMED ontology up and running. Therefore, I wrote a blog post with the different required steps to set it up. I also show how we can semantically enrich data, using example data describing headaches. - The new Smart Flanders API: a demo
Smart Flanders released a new version of their API in October 2017. In this blog post, I show how to use the API with a very easy example: plot the occupancy in different parking stations in Ghent. - Predicting the occupancy of NMBS trains using logs from SpitsGids app (v1.0)
An informal write-up about a basic solution to predict occupancies of future trains of SNCB (Belgian railway company) based on historical data.
- [TheRegister January 2020] Amazing peer-reviewed AI bots that predict premature births were too good to be true: Flawed testing bumped accuracy from 50% to 90%+
- [Medium Blog March 2020] Rookie Data Science Mistake Invalidates a Dozen Medical Studies
- [FutureSummits May 2019] A unique machine learning model to predict preterm birth risk (Dutch)
An update of our work-in-progress on birth risk estimation has been given at FutureSummits 2019, organized by imec. - [DrivenData April 2019] Meet the winners of Sustainable Industry: Rinse Over Run
Me and a friend managed to win the first phase of the two-phase competition called "Sustainable Industry", organized by DrivenData. This is a blogpost by DrivenData themselves. - [VRT NEWS November 2018] Nieuwe datatechnologie volgt prestaties wielrenners op in real-time (Dutch)
An article discussing the results obtained from a two-year project in which parties from both the academic and industrial sector collaborated. The goal of the project was to improve technologies to monitor and assist athletes, during and after training. - [Artsenkrant October 2018] Een 'zwerm' radiologen is slimmer dan de computer (Dutch)
A general article about the application of machine learning in the medical domain. A paragraph in the article handles collaborative research with the Gynaecology Department of the University hospital of Ghent in which we are working towards a predictive model for preterm birth prediction. - [PCMagazine May 2017] How does machine learning work exactly? (Dutch) An article about machine learning, for the broad public (computer enthusiasts).
I gave my opinion on white versus black-box machine learning models. This pdf is how the article was published.
Software
Data Science Competitions
-
Kaggle OpenVaccine: COVID-19 mRNA Vaccine Degradation Prediction
Team with Bram Steenwinckel and MT
Gold medal. 4/1636 (0.001%) -
Kaggle Halite by Two Sigma
Team with Tung M Phung, SolverWorld, robga and Kha Vo
Gold medal. 8/1139 (0.01%) -
Kaggle University of Liverpool - Ion Switching
Team with Zidmie and Kha Vo
In the money: third prize. 3/2618 (0.001%) -
Drivendata Sustainable Industry: Rinse Over Run
Team with Thomas Mortier
First round winners. 1/1217 (0.001%) -
Challenge Semantic Web Challenge on Tabular Data to Knowledge Graph Matching 2019
Team with Bram Steenwinckel
Second prize. 2/17 (6%) -
Kaggle Santa's Workshop Tour 2019
Team with Bram Steenwinckel
Silver medal. 26/1620 (2%) -
Kaggle Google Hash Code Archive (2019) - Photo Slideshow Optimization
2/89 (3%) -
Hash Code Google Hash Code Qualification Round 2020 - Google Books
531/10724 (5%) -
Kaggle Don't Overfit! II
46/2330 (2%) -
Kaggle Ciphertext Challenge II
2/74 (3%) -
Kaggle CareerCon 2019 - Help Navigate Robots
6/1449 (1%)
Funding:
I am funded by a strategical basis-research grant, awarded by the Fonds Wetenschappelijk Onderzoek (FWO). My application number is 1S31417N.Education
- Informatics (Python) (2017 - 2018 - 2019 - 2020): In this course, the first year engineers take their first coding steps by learning to program in Python. I assist different lab sessions: spanning from their first print function, object-oriented programming to implementing their own BCH codes. Moreover, an AI video game bot platform is set up every year, where the students have to write a bot that plays a game.
- System Programming (C/C++) (2017): This course is given to second year computer science engineers and third year electronical engineers. They learn how to program in both C and C++. I assisted in 4 lab sessions, created one lab session myself and corrected it. In this lab session, the students had to create a morse encoder and decoder, by using dynamic programming. Moreover, I corrected the exercise part of the exam.
- Big Data Science (2017): I created one lab session, in which students had to go through the entire machine learning pipeline: loading and processing data, exploratory data analysis, feature extraction, feature selection, model selection and hyper-parameter tuning. The given dataset contained occupancy informations of trains in Belgium. The task was to predict the occupancy (defined by three classes) for future trains.
- MyPlan: A mobile application and data analytics platform towards a healthy lifestyle (Matthias De Brouwer and Sander De Wilde)
- Early identification of sepsis risk through the prediction of positive blood cultures using temporal models in tensorflow (Jan-Alexander Nedee)
- Glycemic control in the intensive care unit: new insights using machine learning (Driek Struyve)
- A study on the impact of domain knowledge incorporation into the pre-processing steps of a machine learning process (Eloïse Piret)
- Athlete monitoring through context-aware training schemes (Lorenz Van Herwaarden)
- CONAMO: Continue & contextbewuste monitoring van het trainingsschema van atleten (Dutch) (Jelle De Bock)
- Gestructureerde data-extractie uit medische notities voor het voorspellen van vroeggeboorte (Dutch) (Jelle Vanhaeverbeke)
- Run4Music: four bright students (3 CS & 1 electrical engineer) created an application that both picks suited music according to your running pace and adjusts this music slightly such that the beats of the song are synchronized with your steps. A tweet about their result went trending on Twitter and one of the four students had the opportunity to present their work on one of the most popular radio stations in Belgium: QMusic.
- Creating a generic platform for AI competitions: four enthusiastic bachelor students helped us to think out an architecture/flow to create a platform for AI competitions, which we host every year in the context of the course Informatics, such that we can relaunch it every year with a different game with minimal effort.
- IEEE Access
- PeerJ Computer Science
- WSDM 2020
Changelog
- February 2021: obtained my PhD!!!
- July 2020: Updated sections + pagination & toc + competitions section
- September 2019: Added more articles & reviewing section
- July 2019: Updated information + bootstrap 4 and some styling
- January 2019: Updated information
- December 2017: Created this website