Bot Contest

Well of course you know! It is Paris, but what if they ask you a differant question about capitals and your bot sayd France.

The thing is that the 10 questions are to sift out the total sh*t bots. All questions have words in them which any robot should be able to react to, what, where, when, how, or(see book of AI). Even if your answer isn't correct, it should be clear that the bot "understood" the type of question, and gives a coherent answer, even if that answer is "I don't know, I'm from Barcelona you know..." If you want your bot to be specialized in say Capitols of the world, music, or AI than you can go deeper with your answers. You are always dependendent on the bot/human you talk to however, so that work could all be for nothing

Something must be wrong with the voting system I asked my friends to vote for Steve, (and whatever other bots they wanted), ranking them whatever they felt the bot should get, (So NO I'm not stacking the scores I did NOT ask for perfect scores, just whatever they thought steve deserved.), and Steve's average is still exactly the same as two weeks ago.. 2.86

Something wrong?

I havent seen the scores move in 3-4 days. Maybe they're getting cemented in?

Lunar- I wouldnt call early bots sh*t bots, as even a simple custom AI can take weeks of programming. We should give credit where credit is due. Not everybody can program an AI. I applaud everyone who tries, no matter how successful they are.

Well I know that each time I use a different browser and use the reload button I keep getting the rankings or the site itself blocked.It is almost enough to make one paranoid. Maybe the fbi is staking me.

Point taken, just wanted to defend the use of the questions, to give a general idea how "advanced" the bot is

FBI or CIA?

gosh maybe it's both..

Maybe Dogh'd's even in on it!

Ok, PF people if something isn't working drop me an e-mail and let me know. Only till I saw the Professor's post about the scores was I aware of it. I think Agentland's automatic refreshing system isn't working but I have used their manuel update just now so it should be updated. I will e-mail Agentland tomorrow.

Chris

Lunar, you are correct. When you have 58 bots we needed a way to narrow the field. You are not going to get the judges to seriously talk to 58 different bots. However, I don't want to focus on the 10 question score along. A chatterbot should be able to hold a conversation as well.

Chris

well Chris I did email you about it. and since then it has gotten worse. I even checked out the source code on that page in both frames and it was fine. I use 4 different browsers, netscape, IE, opera and mozilla. when it isn't working for only me, cause I asked other people if they could get in and or see they did not have any problems, then it gets annoying to say the least.
hopefully what you did will help.