Ever since I started working on the Eccky project my interest in robotic research has increased. Wired has a nice article on a human-robot research project at the Early Childhood Education Center, a preschool attached to UC San Diego. They (the Machine Perception Lab) are trying to make the world's first nondisappointing robot. Eccky has been live now for around 6 weeks, and based on the feedback we receive I can tell that in the end it's not about the physical aspects of the robot. Making it move and appear more human-like is nice, but it's not what makes a successful robot. Creating an emotional connection by (perceived) intelligent interaction is way more important. Our approach for the Eccky chatbot, combined with the enormous amounts of continuous feedback allowing us to improve it, could very well become the first nondisappointing robot. At least that's what our users are telling us. And that's even without adding any physical elements...
"But Movellan also sees this daily routine as part of an odyssey to create the world's first nondisappointing robot. Try as we might, we can't stop measuring bots against the ridiculously high standards that Hollywood has set in our minds. We expect them to be able to follow us around the house, picking up after us, chattering with us like C-3PO, reading our emotions with the accuracy, if not the intent, of HAL. While robots have proved indispensable in narrow kinds of work, like assembly lines, when it comes to interactions with unpredictable, flesh-and-blood humans, they have yet to deliver on the promise of more lifelike responses. That's why Movellan and his team have enlisted this young group of chaos specialists. The researchers videotape each session and use the footage to tweak the bots' software."
Geen opmerkingen:
Een reactie posten