Title : A feature representation learning method for temporal datasets
| Presenter | Ward van Breda |
| Abstract | Predictive modeling of future health states can greatly contribute to more effective health care. Healthcare professionals can for example act in a more proactive way or predictions can drive more automated ways of therapy. However, the task is very challenging. Future developments likely depend on observations in the (recent) past, but how can we capture this history in features to generate accurate predictive models? And what length of history should we consider? We propose a framework that is able to generate patient tailored features from observations of the recent history that maximize predictive performance. For a case study in the domain of depression we find that using this method new data representations can be generated that increase the predictive performance significantly. |
Title : AI for Human Values
| Presenter | Frank van Harmelen |
| Abstract | the media are full not only of stories on the successs of AI, but also with stories about the concerns of unintended consequences of potentially uncontrollable AI. Fortunately, this debate is no longer limited to scaremongering journalists or ill-informed politicians, but is also being taken up by the scientists themselves. I will give a brief overview of a number of recent activities in the scientific community around "AI for Human Values", including Stanford's recent "100 year AI report", the Duch Zwaartekracht consortium on Responsible Data Science, the open letter signed by 15000 AI researchers last year, the debates and activities at the recent European AI Conference, and the concern about the Dutch government's policy on autonomous weapon systems. |