Limitations of Audiovisual Speech on Robots for Second Language Pronunciation Learning

<p>Additional material accompanying the HRI&#39;23 publication found at <a href="https://doi.org/10.1145/3568162.3578633">https://doi.org/10.1145/3568162.3578633</a>.</p> <p>Simulations showing the audiovisual speech performed by Furhat during the second language tutoring experiment.</p> <p>Each video consists of the Furhat saying a Dutch word, followed by saying the Japanese translation twice.</p> <p>Videos M1-M10 correspond to the Matched condition, M11-M20 to Mismatched, and M21-M30 to the Computer-generated lip movement.</p> <p>In the experiment, participants interacted with a physical Furhat, but as it is difficult to make video recordings of the Furhat, simulations are shown here to allow for high-fidelity comparison between conditions.</p>