#EDML workshop on evaluation and experimentation aspects ECML PKDD papers are online at imada.sdu.dk/Research/EDML/… Videos, Q&As in Whova. Thank you everyone Alex Mara Evgeniya Korneva SergejZr Lorenzo Perini feel free to tag (co-)authors Sevvandi K. you were missed

#anomalydetection We propose a notion of #confidence measure that captures how #consistent a model's prediction would be for a test example if the #trainingdata were perturbed. For further information, follow my talk at #ECML! whova.com/embedded/speak… Vincent Vercruyssen Jesse Davis

I'm honored to announce that eventually I won the #OverallEngagementAward at #ecmlpkdd20, getting a lot of votes in multiple days by the #chairs and the organisers! My first year of #PhD couldn't end up any better! #AnomalyDetection #Confidence #StabilityMeasure #Dtai #KULeuven

Congrats to Vincent Vercruyssen for successfully defending his PhD thesis “Designing Anomaly Detection Algorithms that Exploit Flexible Supervision”: bit.ly/3rbmUq7 Welcome to DTAI as a postdoc Vincent!

I'm very proud to be awarded the Gustave Böel-Sofina fellowship for a research stay abroad. It'll be a honor to collaborate with the Multi-source Probabilistic Inference research group of University of Helsinki. Looking forward to being there in 2022! DTAI Computer Science | KU Leuven

🧵Water leaks. Credit card fraud. Irregular heartbeats. What do these have in common? They are all examples of *anomalies* that have substantial consequences if they are not quickly discovered! This #ThesisThread explains how to detect anomalies from data. @DTAI @kuleuven (1/14)

I'm thrilled to announce that my #PhDproposal about measuring and exploiting the #uncertainty in #anomalydetection will be funded by FWO! Thanks to Prof. Jesse Davis, Dr. Vincent Vercruyssen and the whole DTAI team for their support! Computer Science | KU Leuven #kuleuven

Thrilled to announce that our work on machine learning with rejection for vehicle engineering received the best application paper award at ADMA21 🥳 w/ Jesse Davis Wannes Meert Bram Cornelis Siemens Digital Industries Software DTAI (1/6)

If you are curious about #machinelearning with rejection in general, check out our survey on the topic on arXiv.org w/ Lorenzo Perini Dries Van Der Plas Wannes Meert Jesse Davis arxiv.org/abs/2107.11277 (6/6)

Celebrating DTAI’s 5 (!) accepted papers at AAAI 2022. Congrats to everyone!

We are on!🎉 (hard to believe😅) ProbAI — 2025 ✌️ University of Helsinki Finnish Center for AI 🦣 @[email protected] Norwegian Open AI Lab @NTNUnorway

#ProbAI 2022 came to an end, and we truly hope it was a positive experience for everyone. 🙏 Our gratitude goes to the wonderful lecturers, students, general chairs, executive team, teaching assistants, sponsors, ...

Active learning and multi-armed bandit algorithms are a natural fit! 🎰 Read all about it in our #ecmlpkdd2022 research paper: bit.ly/3DIahvp Joint work \w Jesse Davis, Wannes Meert, and Lorenzo Perini.

The ProbAI 2022 materials 📚 and recorded lectures 🎥 are all available to the public. github.com/probabilistica…

Day 2 of #AAAI23 was book ended by morning & evening keynotes. Our AAAI experts discussed artificial intelligence application and processes, as well as the organization as a whole. 📸: rhlphoto.com

⚠️ Exciting news! ⚠️ Our paper on #Bayesian estimation of the #contaminationfactor in #anomalydetection has been accepted to #ICML2023. Check out our arXiv preprint to see how we improve anomaly detectors' performance using a Bayesian approach. doi.org/10.48550/arXiv…

We've extended the deadline for #ECAI23 workshop on verifying learning AI systems to July 3rd. Please consider submitting. w/Luc De Raedt Hendrik Blockeel Bettina Berendt Benoît Frénay JF Raskin Fredrik Heintz dtai.cs.kuleuven.be/events/VeriLea…

Tomorrow (1st October) I will be in Krakow for #ECAI2023 European Association for Artificial Intelligence presenting a preliminary work called "Deep Neural Network Benchmarks for Selective Classification" at the Verilearn workshop. This is a joint work with Lorenzo Perini Jesse Davis and Salvatore Ruggieri. 1/2

Congratulations to DTAI PhD student Lorenzo Perini for successfully defending his PhD thesis on "Operational, Uncertainty-Aware, and Reliable Anomaly Detection" shorturl.at/klTX0 🎓👏🥂🍾

The joint work with Lorenzo Perini Jesse Davis and Salvatore Ruggieri is finally out in the Journal of Data-centric Machine Learning Research. A huge thanks to my three coauthors for their help.