Slashdot Mirror

← Back to Stories (view on slashdot.org)

Interactive, Emotion-Detecting Robot Developed

Posted by Soulskill on from the uncanny-valley dept.

cylonlover writes "A team of Cambridge University researchers have now developed a system that can not only detect a user's emotional state, but can also make expressive responses of its own. Using a robotic likeness of the godfather of the programmable computer, Charles Babbage, the team has hooked the system up to a driving simulator and created a computerized driving companion and navigator that reacts to the driver in much the same way as a human passenger."

7 of 58 comments (clear)

  1. Great by bobstreo · · Score: 2

    Creepy dead-faced robotic backseat drivers.

  2. Has author never used computers? by scdeimos · · Score: 2
    From TFA:

    Voice recognition software is now quite a capable means of entering information into a computer system.

    guffaw.

    Still, where do I get two of these bots from so I can use the T3 lanes on the way to work in the morning?

    1. Re:Has author never used computers? by Dhalka226 · · Score: 4, Interesting

      I have what would be described in science as a "fucking awful" voice. I remember playing with some VR technology a few years back and I couldn't even make it through the training. I adjusted my mic and everything like it asked and then got to the prompts: "Say 'dog.'" "Dog." "You said 'b93r.' Say 'dog.'" After six or seven tries of that kind of nonsense I pretty much gave up.

      I was actually playing with a copy of Dragon Naturally Speaking a month or two ago and was surprised by how good it was. I could tell it was struggling with my voice; the instructions at the start of the training said that most people wouldn't need to read the whole story for the training to be comfortable, and I had to go all the way through and more, but it was pretty good. It could recognize whole sentences at a time, whereas previous technology couldn't even manage words. I finished up the training and busted open Word to give it a try (I could have trained it further but I was just playing with it) and it was very accurate. And this, quite obviously, is with a voice that gives VR technology fits and a general vocabulary where it couldn't even attempt to muff the results by choosing from known word lists.

      I have little trouble believing that with a bit more training and a little more user training on my part (working to enunciate a little better, etc) that it could be a very capable means of entering information. Doubly so if the information I'm entering is predictable such that it has (for lack of a better term) a smaller dictionary to guess from.

      VR is probably pitiful compared to where we thought we would be in 2010 years ago, but it's actually getting to be quite respectable.

  3. Fear by Anonymous Coward · · Score: 2

    How long before it can smell fear?

  4. Re:Why? by oztiks · · Score: 3, Insightful

    We cant even get voice recognition to work because of tonal changes and ascents.

    Recently an article was published talking about the human synapse which describes it far as more complex than original thought. See http://www.medicaldaily.com/news/20101117/3877/new-imaging-method-developed-at-stanford-reveals-stunning-details-of-brain-connections.htm

    I recon we should start focusing creating LtCmdr Data before we progress on figuring out how to build Lore.

  5. Because... by Richard+Kirk · · Score: 2

    I don't think the eventual aim of this work is to have a talking head of Charles Babbage on your computer. The aim is to understand the nature of subliminal communications, and the first stage of doing that is by direct mimicry using an accurate dummy.

    So, what's this got to do with computer interfaces? Well, when most people get annoyed with their computer, they shout at the screen. Why? If the system has locked up, then the error happened in the CPU or the GPU which both sit in the body of the machine unless you have an Apple. Nevertheless, we seem to take the screen of a computer as the 'face' of the machine, and it is the 'natural' target for our gestures and expressions. We are probably making expressions at our computer much of the time. Many laptops have built-in cameras that could pick up these expressions. At the moment, we are doing nothing with this information, but one day we might. The idea that ordinary computers will transmit expressions back at us is a bit more far-fetched, but Japanese companies are looking at using robots for receptionists and ticket sellers, and other jobs where rapid and seamless low-level interaction with a lot of people is needed.

  6. Re:Why? by nospam007 · · Score: 4, Funny

    No, we need hidden sensors in our glasses capable of telling us secret things that we cannot detect in the real world:

    "She's being sarcastic, she didn't really mean that."

    "She's really mad, even if she says she ain't"

    "By 'nothing' she meant you did a major screw-up"

    etc