Slashdot Mirror
Social Robot?
smashr writes "Researchers are currently putting the finishing touches on robots that will be attending the AAAI (American Association for Artificial Intelligence) conference this year as part of the AAAI robot challenge. In addition to robots wearing tuxedos and serving drinks, several robots designed to actually register themselves will be participating in the conference. One such robot is GRACE, being built by Carnegie Mellon University and the Naval Research Lab (among others). GRACE features a digital face and speech recognition to interact with people attending the conference. (She even runs Linux!) Her goal is to register for the conference, give a speech and answer questions. Stories at: Pittsburgh Post-Gazette, CNN.com, and USA Today."
maybe one of these robots could give Bruce Perens' illegal speech! what a setup!!
Cretin - a powerful and flexible CD reencoder
"solicited drama students to teach GRACE how to act like a human so it will make people feel comfortable"
You have to give the builders points for knowing their limitations.
"Lord, grant that I may always be right, for Thou knowest that I am hard to turn" -- A Scots-Irish prayer
Now tell me if I am wrong but, dont you think if an unsocial geek can program a social robot, the geek himself has the capabilities of being social himself?
Being able to program a computer to do something well doesn't mean you can do it yourself. I've written programs that play chess much better than I can.
Another factor here is the "dancing bear" effect. A dancing bear isn't impressive because it dances well. It is impressive because it dances at all. Behaviour that is amazingly social for a robot may be pretty dismal for a human.
A legparnasom tele van angolnaval.
There are exceptions; people are trying to make magical leaps over some percieved barrier, usually self-awareness. The cyc project works along those lines.
By and large, though, AI research occurs in little steps, and most of those steps are like this one. There is a need for a good understanding of what works and what doesn't work for smaller tasks before we're going to solve the bigger problems. What you seem to call "Cognitive Imitation" I would often call "Trying to understand cognition".
In other words, when it comes to "AI", these things tend to be really heavy on the "A" and really light on the "I".
I'd like to hear of an approach that you think is light on the "A" and heavy on the "I".
Recognizing human emotional states and predicting their responses from facial expressions and actions is one of the things humans do best ...<snip>... rather than having them try to do things we're GOOD at.
Bwuahahahahahahahahaa!!!!!
Sorry.. but if we're so good at recognizing emotions and predicting responses from facial expressions, then why are there so many divorces? That alone is proof that we SUCK at recognizing each other's needs..
On that note, maybe I'll start work on AICounselor_v1.0Beta.tar.gz now.. I'm sure GRACE would enjoy that module.
---
With all things being equal.. well, no, that's boring..
Actually, that is planned for the future. We currently have the ability to locate the AAAI badges that are given to all conference attendees and use OCR to extract the name of the person we're talking to (Swarthmore did that portion of the work). Next year, we'd like to try to schmooze with conference attendees about their current research by pulling down their current papers, c.v.'s, etc. off the web. In time, in time. :)