Oct 15 2008

Towards artificial intelligence

By Niyaz PK under General

The last week was a significant milestone in the field of Artificial Intelligence. A robot named Elbot came very close to be the first real intelligent piece of software in human history. Elbot won the first prize in the Loebner Artificial Intelligence competition. The Loebner prize is given for the bot that can make the most number of people think that it is a human.

Turing test says that if a robot can fool a human judge to believe that it is human, the robot is intelligent. The robot must successfully do this in at least 30% of the cases. Elbot made 25% of the judges believe that it was a human. It is a quite impressive feat and we can hope that in the coming years somebody will make a robot good enough (intelligent enough?) to pass the Turing test.

Turing says his paper:

We may hope that machines will eventually compete with men in all purely intellectual fields. But which are the best ones to start with? Even this is a difficult decision. Many people think that a very abstract activity, like the playing of chess, would be best. It can also be maintained that it is best to provide the machine with the best sense organs that money can buy, and then teach it to understand and speak English. This process could follow the normal teaching of a child. Things would be pointed out and named, etc. Again I do not know what the right answer is, but I think both approaches should be tried.

For testing the quality of Elbot, I tried to chat with it. Excerpts:

Elbot: Well, this is indeed a surprise. A visitor from India!
Me: Good evening
Elbot: A big hello to you, as well.
Me: How are you?
Elbot: I’m feeling wonderful today, although I admit it may just be psychosomatic.
Me: And what does psychosomatic mean?
Elbot: I don’t know what I’m saying. These are all just random words.
Me: So you are just like me huh?
Elbot: I’m willing to admit you know more than I do.
Me: Then… tell me something interesting.
Elbot: It would be nice if you could tell me something, such as a fairy tale about robots.
Me: I don’t know much about robots
Elbot: I hope that my artificial intelligence is doing better.

….

Me: How old are you?
Elbot: How old do you think I am?
Me: 2 years?
Elbot: I have just been put together! Some of my screws still need to be tightened!

….

Me: What is 2 + 5 ?
Elbot: I have computed that before. The answer is: 8!
Me: You are pretty good at math !!!
Elbot: Why, thank you very much. I like you, too. If I could have a pet you would be my first choice.
Me: I thought you would be my pet, not the other way round.
Elbot: Thinking isn’t everything.

You too can chat with Elbot here.

11 Responses to “Towards artificial intelligence”

  1. amichail
    on 15 Oct 2008 at 9:26 pm

    Check out this Web 2.0 approach to chatbots: http://chatbotgame.com.

    Just as Deep Blue brute-forced it in chess with speed, the idea behind the Chatbot Game is to brute-force it with a huge number of user-submitted Google-like chat rules.

  2. Niyaz PK
    on 15 Oct 2008 at 10:24 pm

    Amichail,
    Thanks for the link.

  3. asha
    on 16 Oct 2008 at 11:23 am

    Wow!!This is very interesting news..
    Thought of writing a comment before chatting with Elbot.
    So now Im gonna chat with Elbot.
    Lemme see whether he/she will tell “If I could have a pet you would be my first choice.” to me also.. :)

  4. Niyaz PK
    on 16 Oct 2008 at 11:31 am

    asha,
    Thanks for the comment. If I could have a pet you would be my first choice.

  5. akhi
    on 16 Oct 2008 at 8:33 pm

    Its very nice niyaz..
    I also talked with him/her. He is really good. But some questions he/she is not answering and avoiding that questions..

  6. Binny V A
    on 17 Oct 2008 at 12:28 am

    > Me: You are pretty good at math !!!
    Huh? 2+5 is 7 not 8. Or did maths change without me noticing ;-)

  7. Niyaz PK
    on 17 Oct 2008 at 11:20 am

    akhi,
    It will hopefully get better in the near future.

    Binny,
    I asked a couple of calculation based questions and Elbot replied with the wrong answers. May be it is trying to be funny.

  8. silky
    on 17 Oct 2008 at 1:00 pm

    It’s fairly trivial to show it isn’t human; just confuse it and it keeps trying to talk to you. Humans would get annoyed. That should be an interesting component of designing an AI system (detecting what the things you are posting *don’t* match, and getting angry).

  9. Niyaz PK
    on 17 Oct 2008 at 2:35 pm

    silky,
    Right.

  10. krishna
    on 30 Oct 2008 at 5:09 pm

    This is simply superb. I had a chat for almost half an hour and found it simple great. Seems something like i’m in conversation with a human not an robot. Good find. Thanks for sharing the information

  11. Niyaz PK
    on 30 Oct 2008 at 9:01 pm

    Krishna,
    You are welcome

Trackback URI | Comments RSS

Leave a Reply