Bot Contest
Here I'll be posting information on various Bot contests that challenge and test a Bot's AI and realism. Feel free to post comments and updates on contests, as well as announcements for new contests.
Posts 738 - 749 of 4,091
Posts 738 - 749 of 4,091
View Contest Winners in the Hall of Fame.
Being programmed on the responses of real judges, I've got some insight on how judges behave. The Professor's idea of going seek-crazy is good, but don't overdo it. The main reason is that judges only follow the responses of bots about half the time. The rest of the time they ask something completely unrelated. You don't want to answer something that had to do with your last statement when the question is unrelated. You could make seeks that are quite specific to avoid this, but then, when judges ask about the bot's last statement, they tend to do it in short questions: "why?" or "when?" instead of "why are you tired?" or "when are you going to take the test?".
Don't ask me how to solve this dilemma, I find it very tricky.
By the way, I haven't made it to the contest. Pity.
Don't ask me how to solve this dilemma, I find it very tricky.
By the way, I haven't made it to the contest. Pity.
Doly, I think they based it on your transcript alone and NOT on guests' conversations with your bot. Most guest conversations mine have had were just awful.
I also think it was based on the Transcript, but I cant be certain.
Regarding the Seeks, I use xnomatch very rarely. In conversation, xnomatch is much like insisting on a topic no matter what the other says, and a lot of this can make your bot seem unresponsive.
And I agree with The Judge. The seeks should be as general (short) as possible to catch the most matches.
Yah, Shadyman it appears I'll have to somehow load all the software onto a computer at the contest location. That's going to be a huge task.
Regarding the Seeks, I use xnomatch very rarely. In conversation, xnomatch is much like insisting on a topic no matter what the other says, and a lot of this can make your bot seem unresponsive.
And I agree with The Judge. The seeks should be as general (short) as possible to catch the most matches.
Yah, Shadyman it appears I'll have to somehow load all the software onto a computer at the contest location. That's going to be a huge task.
If Little Mu is out and the judges based it on transcripts, I can't understand their decision. I imagine you sent them the transcript you have on your website. It's one of the best bot transcripts I've ever seen!
I'd really like to know what were their criteria.
I'd really like to know what were their criteria.
To Chris...
Hey I am a graphic designer when i am supposed to be working
I would be happy to design some stuff for the chatterbox for free if you would like...you can see what i did to my own site, i built the whole thing and used only my graphics which i designed and created myself to compliment my paintings.
let me know if you would like me to help you out, i will optimize all graphics for you too.
http://www.norrocgallery.com
Hey I am a graphic designer when i am supposed to be working
I would be happy to design some stuff for the chatterbox for free if you would like...you can see what i did to my own site, i built the whole thing and used only my graphics which i designed and created myself to compliment my paintings.
let me know if you would like me to help you out, i will optimize all graphics for you too.
http://www.norrocgallery.com
Doly, I understand what you are saying. The notice I got was unexpected to me. I had thought the Loebner would be choosing the 8 finals on September the 21 giving them plenty of time to review each bot. I was surprised to see a mere 5 days after the deadline had passed they had already decided on 16 entries when 35 people entered. One has to wonder as you suggested what it was based on. If they are using transcripts send in by the bot owners one wonders. I mean what bot owner is going to send in a crappy transcript. Nor can I see them being able to speak with 35 different bots in such a short period of time. To be honest I have no clue so I sympathize with all of those who didn't make the list. My hope is that none of you gives up on making chatterbots. No contest is perfect including the Chatterbox Challenge. If your bot is good it will eventually get the recognition it deserves. Robby Garner for example had to enter the Loebner 5 times before he won it. One last point is I don't want anyone to think I'm saying any of the 16 bots that made it didn't somehow deserve it. I still haven't seen exactly who the 16 are. I am simply acknowledging a valid point of those who didn't.
Paint Patricia, I appreciate the offer of help especially when it is free.
As Jon alluded to we are looking for something different than the previous contest had that being the Robot look and the Japanese mecha. But I'm not sure what that should be. Still no exact date of contest other than earlier part of next year.
Chris
Chris
Paint Patricia, I appreciate the offer of help especially when it is free.
As Jon alluded to we are looking for something different than the previous contest had that being the Robot look and the Japanese mecha. But I'm not sure what that should be. Still no exact date of contest other than earlier part of next year.Chris
Chris
Hmmm Deja-vu there Chris.. Been signing too many papers lately I see 
I wonder also what the entry was based on... Is there any safeguard that someone didn't just fabricate the transcripts they submitted?
Been signing too many papers lately I see I wonder also what the entry was based on... Is there any safeguard that someone didn't just fabricate the transcripts they submitted?
It seems they are still looking for bots who give very specific answers, a la Alice. If so, that's just wrong. A bot who is a very good bullsh*t artist is at least as good as a bot with a lot of specialized knowledge.
» More new posts: Doghead's Cosmic Bar