At this year's Interspeech2021, our colleagues Silvan Mertes, Pol van Rijn, and Dominik Schiller were involved with three papers that were accepted for presentation:

• Analysis by Synthesis: Using an Expressive TTS Model as Feature Extractor for Paralinguistic Speech Classification.
• Exploring Emotional Prototypes in a High Dimensional TTS Latent Space
• A Prototypical Network Approach for Evaluating Generated Emotional Speech

The trio also placed second in the conference's Computational Paralinguistics Challenge in the escalation detection subtask.

We sincerely congratulate them on this success.