JADS Alumni Event
On the 27th of November JADS is organizing another Alumni event for Professional Education, Master and PDEng Alumni! We hope you can all join us online to socialize together while enjoying interesting presentations from some of our professors! The program of our event will be as followed:
Program Alumni Event
Program Alumni Event
- 16.00 Welcome by the Chairman of the Day: Emile Aarts
- 16.15 JADS Talk by Ksenia Podoynitsyna about NLP and Text Mining
- 16.45 Switch/short break to get coffee or a beer
- 16.55 JADS Talk by Martijn Willemsen about Explainable & Interactive AI
- 16.25 End of the talks, introduction platform Gather.town
- 17.30 Time to catch up with fellow alumni in social platform Gather.town
Speakers & Content of the day
Ksenia Podoynitsyna about NLP and Text Mining
Ksenia Podoynitsyna is an Associate Professor of Data-Driven Entrepreneurship, serving as the Program Director of the master “Data Science and Entrepreneurship” and Data Entrepreneurship unit head at JADS. Ksenia will provide an overview of recent projects employing text mining and natural language processing techniques in JADS and discuss the overall learning points from these projects. She will also share the recent developments in JADS in general and the master “Data Science and Entrepreneurship” in particular, highlighting the opportunities for collaboration on both education and applied research side.
Martijn Willemsen about Explainable & interactive AI
Martijn Willemsen is an Associate professor of Human Decision Making in Interactive Systems. As head of the recommender lab at JADS, he focuses his research on the user-centric aspects of recommender systems. In his research he focuses on how recommender systems can help people to explore new preferences and tastes, rather than keeping them in their bubble. He develops music recommenders for genre exploration and for groups rather than for individuals. He also has worked on recommenders in the domain of personalized energy saving advice and healthy lifestyles.
In his talk he will discuss the recent interest in explainable AI and tools for interpretability and visualizations such as SHAP and lime. Research shows that data scientists have a hard time to understand the outcomes of their models even when provided with these types of tools which calls for more research on human-algorithm interaction. Users should be able to interact and collaborate with the algorithms, to indicate when a prediction is wrong, or to collaboratively come to a better prediction or advice. He will illustrate this on his recent research on pacing advice for marathon runners to run a personal best that shows collaboration between coach and algorithm could improve model predictions.