Selma Yilmazyildiz Kayaarma
Politeness Detection in Speech for Human-Computer Interaction
We’ll demonstrate our deep learning-based politeness detection system implemented in a smart speaker interaction scenario using an open-source smart speaker (Mycroft Mark1). The users will make various verbal requests (e.g. asking time, weather...) both politely and impolitely. The request will be analyzed in real-time by our deep learning based politeness detection architecture and the response of the voice assistant is customized based on the output of the politeness detection architecture.
Selma Yilmazyildiz Kayaarma is a postdoctoral researcherat the Audio-Visual Signal Processing laboratory of Vrije Universiteit Brussels. Her research interests include social signal processing, human–robot interaction, emotion in speech, affective computing. Her membership list includes the Institute of Electrical and Electronics Engineers (IEEE), the Association for the Advancement of Affective Computing (AAAC), and the International Speech Communication Association (ISCA).