I just want to interject here that replies/rebuttals are welcomed and encouraged as long as they remain polite, and don’t devolve into personal attacks. If this thread even hints at becoming a flame-fest, I’ll close it immediately, remove the offending post, and consider taking further, more aggressive actions.

Don’t make me put on my Moderator’s hat!

Just thought I would add my support to Steve, I agree, these two contests are a sham.  There ends the matter for me.

Dave Morton is

the moderator

http://ai.r-i-software.com/images/moderator.jpg

Haha, very good Vincent :D

ZOMG, TOO funny.

If the image were any larger, I’d “swipe” the head portion and use it as my avatar for a while.

In this small AI/Chatbot community, among the various forums and contests, individual people each have a reputation that precedes them.  Some are known for going out of their way to share knowledge, and to be helpful. They freely give away vast amounts of their work, making it public, answering questions, demonstrating a sense of openness, and projecting a polite and friendly disposition.

Then, there are others known for being difficult.

When I first came across the chatbot community, I was fortunate enough to encounter an individual who was of the former persuasion.  He was inviting, and he had a little chatbot forum, and a core group of people there who were just like him.  And, that’s why I’ve stayed with Pandorabots, and why I’ve made the AI Nexus Forum my home.  It’s also where I met Steve who, more than anyone, has helped me enjoy chatbots.

By the way, Knyte Trypper was once banned from the RAB forum and wore it as a badge of honor.  I wonder how they’re doing these days?

http://knytetrypper.proboards.com/thread/266/knytetrypper-banned-rab

Steve Worswick - Jun 18, 2013:

Wendell/Chris - If the only way you can win a contest is to run your own, I suggest improving your bot instead of faking a win.

I am not a poor loser and don’t care about winning, as long as I see a contest that is run fair and transparently.

That oddly similar flash video biz really made me wonder about the connection between these “contests”.  I suppose even if the judging scores seem fair, the real meat (in getting a higher score) is in the exact phrasing of the questions, which can easilly be framed to favor one bot over others.

But how to overcome this, even if incidental, issue?  Where can unbaised questions be derived from?

But how to overcome this, even if incidental, issue?  Where can unbaised questions be derived from?
Carl
Another member and I (im not going to use a name since I havent asked him about posting his name in conjunction with this, but he can jump in if he feels like it) have been throwing some ideas around regarding this very question. THe idea that we have sort of agreew that makes since is this.
1 Each botmaster submits a list of ten questions
2 A moderator takes these questions and using an app (psudo random selection doo hickie) picks one question from every botmaster. The number of questions asked will necessarily equal the number of bots signed up. That way every botmaster knows that there is at least one question that his bot will excel at, and everyone knows that there is no chance that any one bot (and by the way Im not saying that is or has ever been the case) is favored over any others. The app will be made available for download, but really everyone participating should be able to readily see if there is a problem. Either one of your questions is included or it isnt.
3 Instead of single category numeric scores, there will be categories.
Most logical
Most entertaining
etc

3 In addition to the question\statement segment, each bot will be expected to run freestyle for 5 minutes. The judge freestyling with the bot will be expected to tailor the topic to the bots personality or page elements. Things that the bot, in being what the bot claims to be, would be able to converse freely about.

4. in addition to the judges scores, judging will factor in public polls (on facebook) Havent worked out the math on that yet, but maybe 50\50 will do it. To ensure that the polls are not “jacked” each transcript will be on a separate active page that will count unique page views, and this will be compared against the poll. If 10 people read the transcript and 10,000 vote, we can be pretty sure that the poll does not reflect the number of actual views.

Ive started to out together a site that might run something just for fun, maybe until Donald gets his league up

http://ai.r-i-software.com/bragging_rights/

If anyone has any suggestions or sees some glaring flaws please jump in

Vince

(By the way the “giger headressed” alien\cyborg girl is my oldest daughter. Ive been dying to use that photo in case anyone was wondering where the “theme” came from LOL)

Thanks for your kind words guys both on here and via email. His banning me really doesn’t matter though. I had no intention of entering.

If you want a laugh, pop over to http://www.daniellechuchran.com/phpbb/viewtopic.php?f=13&t=782
It looks like he is making alter egos to have a further go at me.

I still don’t know why he thinks I tried to enter his contest though. If that is the entry form at the bottom,  I have no idea who Martin Gould is?

So that’s what Dave looks like in real life! Nice work Vince.

Steve was the reason I got interested in AIML. Although I was into other technologies at the time, it seems a given that knowing something about AIML was useful. So I went about creating a simple bot. It didn’t grab me. I was too stuck on Verbots and felt limited. Although that was not so much the fault of AIML, seeing the extras Steve provided really grabbed my interest, and took me further into AIML than I would have gone at that time.

—-

On a different note, what some people may not know, and I hope Steve doesn’t mind me saying, is that he was one of a few people responsible for saving CBC one year. And he may not know that I know that.

Basically, through the poor organisation and total-control approach of WC (that’s at least how it seemed) the forums were buzzing with complaints. Not simply flaming. Genuine, serious complaints, with real foundation. That year’s competition almost came to a stop.

Steve and a few other people were asked (by WC) to help, and agreed. That small group put a lot of time and effort in a short time to save that competition that year.

The irony is that CBC may not have survived if the negative effects of that year had not been reversed, which was down to Steve and that small group.

WC has conveniently forgetton this it seems.

Vincent Gilbert - Jun 18, 2013:

4. in addition to the judges scores, judging will factor in public polls (on facebook) Havent worked out the math on that yet, but maybe 50\50 will do it. To ensure that the polls are not “jacked” each transcript will be on a separate active page that will count unique page views, and this will be compared against the poll. If 10 people read the transcript and 10,000 vote, we can be pretty sure that the poll does not reflect the number of actual views.

Most of it sounds pretty good, but #4 is probematic- Facebook “likes” are not a reliable way to gauge bot performance (“like” links could and would be put anywhere and everywhere). Maybe only registered forum users could vote, or something similar.

Most of it sounds pretty good, but #4 is probematic- Facebook “likes” are not a reliable way to gauge bot performance (“like” links could and would be put anywhere and everywhere). Maybe only registered forum users could vote, or something similar.

Carl,
I agree that “likes” would be misused. Facebook has a feature that allows you to create polls (basically just question\option lists) that will automatically tabulate data. These can be “shared” but unlike “likes” I do not believe that they can be accessed from anywhere other than Facebook. Doesnt mean that it couldnt be problematic though, I have just started looking into it.

On other aspects of this thread, again Im not commenting on a contest or promoter but I feel that in overall terms it is a gross misrepresentation to refer to Mitsuku as an Alice clone.  And by that I do not mean to imply anything negative about ALICE either, far from it. But as I understand it the term ALICE clone refers to a bot that has downloaded the base ALICE set and then done little or no work furthering the personality.  I visit Mitsuku myself from time to time and I am still astonished by her responses.

Vince

Vincent Gilbert - Jun 19, 2013:

... it is a gross misrepresentation to refer to Mitsuku as an Alice clone…

To be honest Vince, I thought this was cleared up about 3 years ago and I’m not really sure why WC has decided to dredge it up again. Mitsuku came top of round 1 of the 2010 CBC (even beating the original ALICE which I was supposed to have cloned from). However, one person then complained about Mitsuku, as it had beaten his bot. Nobody else was bothered. You can probably guess who this person was.

When I read that Steve Worswick and his Mitsuku were banned permanently from the Chatterbox Challenge, I was a bit distressed.

Maybe it would be better to write this in the CBC forum, but as I don’t have put a bot into that contest, I can’t apply to that forum, so I have to write somewhere else.

I’m trying very hard to understand both sides, but as on the CBC forum, there is mostly Steve bashing, so it is not easy to do so.

First, iirc, the CBC rules used to state that no blatant ALICE clones had the right to enter. It was rather obvious that time that Mitsuku is based on ALICE’s database, but I wouldn’t call her a blatant clone. I remember that in 2010 I thought she wasn’t imho, but it was justifiable to think so., so I thought it’s a pity but not a shame that Wendell decided to disqualify her. (But isn’t English Tutor, who took part in 2010 too, based on ALICE too?)

That Steve asked what the participants’ money would be used for, can be understood as sarcasm (and even as an insult), but it can be an honest question as well; this is not a thing I can decide.

That Bildgesmythe overtook Mitsuku in 2011 and 2012, I’d call bad luck for Mitsuku, no more, no less, until there is positive proof the CBC question were made known to Patti beforehand and/or worded in a way that Bildgesmythe would be sure to win.

I didn’t think much about the similarity of the Flash movies, for I know there are tools to simplify making them giving similar results every time. The “verse/verses” instead of “versus” error can hint there is a close connection between those two people, but never having learned Latin plus attending to the same forums would suffice.

In spite of all that (near-) flaming, I don’t give up hope that - with a bit of indulgence, forgiveness, and common sense, and maybe a bit more - a good relationship between Wendell and Steve can be restored.

Whatever it is, there is no doubt that Mitsuku is one of the best chatbots on the Internet, and that Wendell is doing a really, really great job with the CBC.

Peter Wolff - Jun 19, 2013:

In spite of all that (near-) flaming, I don’t give up hope that - with a bit of indulgence, forgiveness, and common sense, and maybe a bit more - a good relationship between Wendell and Steve can be restored.

Whatever it is, there is no doubt that Mitsuku is one of the best chatbots on the Internet, and that Wendell is doing a really, really great job with the CBC.

Wise and thoughtful words indeed.