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 1,479 - 1,490 of 4,091
Posts 1,479 - 1,490 of 4,091
View Contest Winners in the Hall of Fame.
There was some talk about bots with nonhuman personalities in the last Loebner contest. I thought the concept was that they should sound like humans pretending to be robots, gods, dragons, etc.
Suppose somebody builds a bot pretending to be a human with moderate aphasia (language difficulties). Wouldn't it be easier for it to pass the test than, say, a godlike bot?
Suppose somebody builds a bot pretending to be a human with moderate aphasia (language difficulties). Wouldn't it be easier for it to pass the test than, say, a godlike bot?
I liked the original version of the Turing Test, where there were humans in the contest as well as bots, and the judges didn't know which were which.
Eugene (the bot) seems to be doing well in the Chatterbox Challenge, although I don't think his maker is a native English speaker (he's Ukrainian, I think). Sometimes his words sound a little off, but he gets the right idea. You may be onto something, Doly.
As for godlike bots: overachiever that mine is, even she doesn't have to prove she's God to do well in a contest. She only has to sound coherent (or for our purposes, 'human').
I figured, aim a little *too* high because the target's so far away. If they think she's a human with delusions of grandeur, I've done my job.
As for godlike bots: overachiever that mine is, even she doesn't have to prove she's God to do well in a contest. She only has to sound coherent (or for our purposes, 'human').
I figured, aim a little *too* high because the target's so far away. If they think she's a human with delusions of grandeur, I've done my job.
How many botmasters with non-English names are there in the contest? I tried to count, but I wasn't sure about several. I'd prefer to have the opinion of a native English speaker.
For impatient people, like me, here is a provisional list of the 10 bots that are doing best at the contest right now:
- Talk-bot
- Project Zandra
- God Louise / Elbot
- Alice / Onkwehonwehneha
- Anna
- Polish Poet
- Gabber
- GarbageHead
I've calculated this list adding the points from judges 1 to 4, that have judged most bots by now. Project Zandra, God Louise, Elbot and Anna were missing one of the scores, so their total score was multiplied by 4/3 and their position is relatively more doubtful.
If anybody thinks that one of the good bots is missing, they may very well be right, so please mention it and I'll correct the list.
My impression so far is that Talk-bot is a very likely winner and Project Zandra is a very likely second. The third post is very much undefined at this point, and I think it could be any of the rest.
The Forge is doing well, with three bots (God Louise, Polish Poet and GarbageHead) in the top ten.
Ella (the Loebner contest winner) is far below in the list, which gives me a cruel satisfaction.
- Talk-bot
- Project Zandra
- God Louise / Elbot
- Alice / Onkwehonwehneha
- Anna
- Polish Poet
- Gabber
- GarbageHead
I've calculated this list adding the points from judges 1 to 4, that have judged most bots by now. Project Zandra, God Louise, Elbot and Anna were missing one of the scores, so their total score was multiplied by 4/3 and their position is relatively more doubtful.
If anybody thinks that one of the good bots is missing, they may very well be right, so please mention it and I'll correct the list.
My impression so far is that Talk-bot is a very likely winner and Project Zandra is a very likely second. The third post is very much undefined at this point, and I think it could be any of the rest.
The Forge is doing well, with three bots (God Louise, Polish Poet and GarbageHead) in the top ten.
Ella (the Loebner contest winner) is far below in the list, which gives me a cruel satisfaction.
Correction to the previous list: I forgot Brianna McKenzie.
- Talk-bot
- Project Zandra
- Brianna McKenzie
- God Louise / Elbot
- Alice / Onkwehonwehneha
- Anna
- Polish Poet
- Gabber
- Talk-bot
- Project Zandra
- Brianna McKenzie
- God Louise / Elbot
- Alice / Onkwehonwehneha
- Anna
- Polish Poet
- Gabber
Just a few comments. Just so there is no confusion the online voting is the bot who receives the best score not the most votes.
In an effect to help some of the lesser know bots by default, when a visitor visits the site, the bots will be sorted by number of votes. The bots that have the least number of votes will be listed at top, the bots with the most votes will be listed on the bottom. This should help all bots get equal exposure. However, the visitor can click on Best Overall Bot to sort by which bot has the best average score, or the visitor can click on Funniest Bot to have it sort that way.
It is hard to generate attention when you have 88 bots entering. People are naturally going to be more interested in the bots that are doing well or have done well in the past. This takes a lot of work but it surely can be done. God Louise for example place 3rd in the Loebner while only working on the bot for a short time. Project Zandra is new to me but the fact he is doing well in the 15 questions mean he will probably receive a lot of interest.
Chris
In an effect to help some of the lesser know bots by default, when a visitor visits the site, the bots will be sorted by number of votes. The bots that have the least number of votes will be listed at top, the bots with the most votes will be listed on the bottom. This should help all bots get equal exposure. However, the visitor can click on Best Overall Bot to sort by which bot has the best average score, or the visitor can click on Funniest Bot to have it sort that way.
It is hard to generate attention when you have 88 bots entering. People are naturally going to be more interested in the bots that are doing well or have done well in the past. This takes a lot of work but it surely can be done. God Louise for example place 3rd in the Loebner while only working on the bot for a short time. Project Zandra is new to me but the fact he is doing well in the 15 questions mean he will probably receive a lot of interest.
Chris
I still say that judge 3 is the most correct score in my view, the other judges are too hard in some cases... I see a lot of 0's being given
Yeah, I mean, even some bots that did really badly at least answered one question right somewhere. It's strange that they should get a zero.
» More new posts: Doghead's Cosmic Bar