listening to reinforcement learning

Scurto et al., 2018
psychoacoustics, computer science

A quantitative study on how people perceive AI behaviours through sound. We invited participants to guide three reinforcement learning agents in two sound synthesis environment using positive or negative feedback. One agent was always exploring, one always exploiting, and one balancing exploiration with exploitation. Participants successfully interacted with all three agents to reach a sonic goal. Subjective evaluations showed that agents’ exploration behaviour, rather than their learning to reach a sonic goal, was critical to how participants heard them as collaborative.

Preliminary study for the Co-Explorer.
Year
2018
Credits
The project was developed with Frédéric Bevilacqua and Baptiste Caramiaux in collaboration with the ISMM group of IRCAM and the ex)situ group of LRI (INRIA), in the context of a PhD thesis at Sorbonne Université.
Publication
Paper at SMC (2018)

This slideshow requires JavaScript.