Posted by: Guillaume Faure 3 years, 7 months ago
Abstract :
ECHO is a small booth in which the visitor takes a seat and sees his face reflected as if there was a mirror but it’s an optical effect. The machine records images and sounds continuously, and then uses the collected words to respond to the visitor who started a conversation with his reflection. When ECHO speaks, the video stream is replaced with the video sample that corresponds to the recorded sound, giving the feeling that the reflection answers to the participant.
This AI repeats the words of the visitors in a shifted way according to the criteria of the tone’s harmony of the voice. Its speech is fluid in the conversation in terms of musicality, but the meaning of words is "random". The continuity of ideas in the discussion and the natural need to find meaning allow the participants to follow a conversation, sometimes made of embarrassment, laughter, clarity, or confusion...
This prototype let us spend a moment with ourselves. At first it confronts us with our face and voice and then the interaction quickly arouses curiosity, and a dialogue is established. By repeating and shifting in time sentences, the device puts our words out of context.
The first thing we perceive about a language is its music, which means its prosodic organization.The rhythm and pitch tone of the voice and silences are the input factors for the AI, not the meaning or context of the words. That’s why ECHO can adapt to any language.
Share on Twitter Share on Facebook
Comments
There are currently no comments
New Comment