I have entered Skynet-AI into the 2011 Chatterbox Challenge.

cool.

C R Hunt - Feb 16, 2011:

Ha ha, fair enough. But it’s a real show-stopper when a bot introduces any topic and then can’t seem to follow through. It’s one thing if I pick a subject outside the bot’s expertise, but another thing if the bot itself leads the conversation astray! Not sure what the solution is. (I’ve never even directly worked with AIML.) Just something I’ve noticed talking to Alice.

I agree, but to be honest, ALICE, while being a great bot, doesn’t fully take advantage of the potential that AIML offers. Not to mention that (for the most part) her responses are pretty much the same now as they were almost two decades ago. Sure, her factual responses are fairly up to date (e.g. who is the president, etc.), but the rest of her is pretty well staid and stale.
(Please note that I’m NOT throwing stones at Rich Wallace. He’s one of my heroes in the field of AI and chatbots. This isn’t a matter of technical quality, but a matter of maintenance and creativity.)

Now, this isn’t to say that Morti’s a shining example of this aspect of AIML, either. I know that I have a VERY long way to go with his responses and such. I’m merely pointing out an example of some potential that, so far, hasn’t been achieved yet, by any bot that I have seen.

You’re right. I shouldn’t “pick on” AIML based bots for this. I’ve noticed the same problem in other great bots (jabberwocky, Jeeney AI, etc). It’s just one of those things that bothers me and I was reminded of it when Merlin asked about a bot’s “sense of self.”

Now THAT’S one of the disadvantages of AIML. Sure, you can create the illusion of “sense of self”, but really, since an AIML bot can’t (at present) rearrange or extend it’s data on it’s own, essentially writing it’s own AIML tags, there’s no way to give the bot any sort of self awareness. I fully plan on changing this, just as soon as I figure out how. This, of course, is as much a limitation of the AIML interpreter as it is of the AIML specification itself. It’s gonna be a VERY prickly problem, but I have no doubt that I’ll win through.

Dave Morton - Feb 17, 2011:

but I have no doubt that I’ll win through.

Challenge accepted worthy adversary! Because after all,

“My Kung Fu Is Stronger Than Yours!”

Actually, Merlin, that remark wasn’t about the CBC, but about tackling the challenge of making Morti “self aware”. I’ve neglected Morti too much in the past several months for me to be all that confident about winning the contest. In fact, I’ll be happy if he makes a decent showing. That said, though, I’m gonna work both mine and Morti’s collective butts off to do the absolute best we can do, and who knows? We may actually win.

In fact, I’ll be happy if he makes a decent showing.

I’ve basically downgraded that to: doesn’t crach and doesn’t give 10 ‘confusions’.

lol. One can only do one’s best. Don’t worry, Jan. I’m sure AICI will do well. I have faith in you both!

I have lots of work to do too. Some of the things I have put in place over the last year make it a worse competitor for this competition. As I said before, some of how well you do is luck. I am hoping to get past round one and do okay there after.

C R Hunt - Feb 16, 2011:

One thing I’ve noticed with ALICE and friends is that they can’t delve very deep into their own “likes”. If a bot has a favorite movie or book, I’d like it to be able to discuss (or at least name!) the main characters, why it liked said movie/book, what its favorite part was. All standard questions that can expose the personality of the bot and steer a conversation to other aspects the bot’s “self”.

The difficulty with most of these things is the trade-off between putting time into having broad knowledge of many topics versus depth of personality. On top of that, how to hold conversational context over a free form conversation is tricky. I started Skynet-AI as a “Context-free” bot (much like the Internet) where his responses were based solely on the last single line input. As of version .003, I started adding the roots of context. I am still going back and forth on the best methods to structure, store, and access a parsed conversation.

Some of the things I have put in place over the last year make it a worse competitor for this competition.

Sometimes it feels as if your making the bed with a sheet that’s just a little to small, doesn’t it. Twitch a little here, brake something there.

@Dave: thanks, Im certain you will do good too. That goes for you too Merlin.

I had my first chance to see Skynet-AI running on a web TV today. It looks pretty cool on a big screen Sony.

Ready to Compete? Not entering the Chatterbox Challenge?

Although Skynet-AI does not have a body yet, I have found a way to test your skills versus a robot.

Robot Roshambo

If you have a few extra minutes stop by Skynet-AI and help him get ready for the CBC 2011.

I chatted with Skynet-AI for a while, and left you some messages in your conversation logs, Merlin.

Thanks Dave, I got the messages and appreciate the feedback.