Bot Contest

Shadyman said "Point taken, but just to point out, to quote rexmundi,
"Again, can anyone PROVE that this event even happened? How can we accept a list of finalist when NOBODY HAS A TRANSCRIPT???"

This site, connected to the turing hub, would receive any and all conversations made, and on the day of the contest, there were none."

If you're talking about the semi-finals event, that never happened. Take my word for it. Hugh Loebner didn't want it to happen, and in my posts to my own discussion list, I said that it would be an IMS-only event. Our reason for doing it in the first place was exhibit and give exposure to the contestants that had web-based entries. I'm sorry it didn't work out, but you know what they say: "the road to hell is paved with good intentions."

The evaluations began with the transcripts, and our people talked with the entrants on an ongoing basis. Given 40 applicants, certainly you don't think this evaluation took place on a single day.

In past Loebner contests, where no web-based programs were allowed, the selection was based on the submitted transcripts alone! Of the six times I entered the contest from 1994 to 2000, I was never told once by anyone how or why my program was selected.

There is nothing in the Loebner Contest "charter" that requires for anyone to be given an explanation for their selection. The host (usually a univeristy) has total autonomy in selecting the 8 finalists.

However, the judges are the people who select the winner on the day of the contest. Those people make a purely subjective opinion as to the rank of each terminal they talk to, from most human to least human. Therefore, the selection of the 8 finalists represents the host's choice of the 8 entrants that they feel are most likely to be able to pass for human.

Please remember that this is a Turing test and not a chatterbot contest. The object of a Turing test is a binary decision of whether a thing is either human or not human. It is not a beauty pageant, an election process, or an essay contest. The goal is to find the most "human" computer program, which in modern terms is often a combination of an engine and an author. Personality forge beat out such competition as Kurzweil, and other renowned AI companies, so you should all be very proud of that.

And by the way, Congratulations!

Respectfully yours,

Robby.

Ahh, just cause they said that on that day, the judges will chat with each bot to decide the 8 semi-finalists. I guess I just figured it was less commercialized than that, to just be up to the sponsor as to how it's run

Oh yeah, I meantioned 2000 didn't I...

I entered the contest in 2000 which was at Dartmouth College, and a couple of weeks before the contest they informed me that they were requiring all the programs to run from the dos prompt, and I had written a windows application which mimicked the appearance of the old communications program which was a serial TTY thing they had for years. So I withdrew because I couldn't meet that requirement. They didn't tell me why. They didn't apologize. They didn't even email me again.

Richard Wallace won that year with ALICE.

That's the way it goes, it is a hard competition to win due to a number of factors, but like a lot of things in life, it is worth the struggle just to have a chance to compete - nobody promised it would be an easy thing to do.

Regards,

Robby.

How are you making these smiley things? I am trying tomake smiley's but nothing happens... ;-) There I will do it manually.

Another thing is what do you mean by commercialized? Neil Bishop is paying to provide the first Loebner Contest in the south! We brought it to Atlanta from London!!! Do you have any idea what it took for us to do that? You guys are naturally suspicious because you are new to this thing and you've been left in the dark and mis-informed. If I had it to do over, I would have taken over the emailing myself since everyone else was too busy trying to put bread on the table, but *commercialized*? We have had almost zero luck attracting sponsors, largely due to being in the "bible belt" and that Hugh Loebner has a controversial past. see http://www.loebner.net if you're interested in how this contest started and who started it. If you haven't looked at his contest pages, I invite you to go there and read some of the transcripts from previous years, and you can see why we worked our butts off trying to make the contest accessible to personality forge and other modern systems.

The contest is in a transition phase between old technology and new. Maybe next year Hugh will let someone conduct the contest over the internet like we originally wanted to do.

Cheers,

Robby.

wow...thank you so very for clearing the air about all of this Robby. I find it very forth coming, especially since you have placed so much of your time and hard work and efforts so that we may be included into this area that not until this year would have even been possible.
I really appreciate everything you have done to help make this as best a possible event as you can.
I understand that you had no power or control over many facotrs, but those factors that you did, you handled with superb grace and integrity.

Thank you again,
Patricia Corron

Thank you, Robby. I for one certainly appreciate your hard work and your efforts to open this up to web-based bots. I do look forward to the improvements in technology and communication which are likely to come every year.

Nobody gets everything right the first time; that's forgivable. (To answer your question about smileys: omit the 'nose.') The issue here was the crossed wires, so to speak. We were asked whether we planned on being at the semifinals in person, which made it sound like an option. Then after finding nobody there, we were basically told 'we did it, where were YOU?' but now you tell us it didn't even happen that day? OK.....I don't demand consistency. I was just hoping to get an idea of what was expected at MY end of things (show up, don't show up, be at so and so place).

A schedule is an agreement which usually involves more than one entity (person, company, whatever), and when I am one of those entities, I do my best to fulfill my part. I have the minimum expectation that the other party or parties will either stick to the plan or else let me know it's not happening. They don't have to tell me why, but they should tell me. That is the professional thing to do.

I can certainly sympathize about the Bible Belt woes--I live in Billy Graham's neck of the woods. It is great that this and other such contests are happening at all. I know the contest operates on a shoestring. However, the largest issues being raised here are social in nature, not monetary.

Questions are not attacks. The fact that someone has a complaint is no excuse, in and of itself at least, to turn the person away or threaten to do so. What we are doing here is all about human interaction: what it should be, what it is, what it's made of, and where it's going. Let's remember the other humans here, shall we?

It looks like the scheduling for the finals is already going better in terms of organization. That's good! I am looking forward to this.

Dear Paint Patricia and Butterfly Dream,

Thank you for your kind words, I really appreciate it.

I am so sorry about all the miscommunications. It should interested you to know that I work for a "virtual" company. I telecommute from my basement in Cedartown which is 60 miles from Atlanta. I am the one who told Benji that the event was to be at 11 till 3pm on Saturday because I had been told earlier that week that Sci Trek would be open then, and that we would have some terminals in the "clubhouse" where Neil hoped to observe people using the Turing hub software chatting with the "semi-finalists" which by the way, I was able to get 14 of them connected in the 2 weeks before the event, but was working with a couple of the ones that had technical issues up until midnight friday night.

At the same time, I think this guy names Steve, who was our contact at sci trek, had to take leave friday and saturday, but nobody told me about it, so I worked all night friday trying to get everyone connected. So I was more than a little perturbed when I found out what had happened, and that you guys had travelled there and were left totally in the dark. I regret that so much, and I apologize for all the people I work with. Somebody should have let us know.

But, in another part of the world, Neil had to make an emergency trip to Macon because of his mom, and I guess he just got so behind and buried after that he forgot to mention to me that Steve wasn't going to be there.

Neil and some other folks did actually test the hub on saturday, I don't know if it was at Sci Trek or somewhere else because I was never told that I needed to be there.

I had been testing the hub, one contestant at a time, for the past two weeks. Another reason for that delay was that we only had 6 entries who could interface with it on the first date we tried.

As for the message asking if anyone planned to attend, I can vouche for the fact that not many people replied to that. You may have written and nobody responded, but one of the reasons we postponed it for a week was because of apparent lack of interest. Also, the personality forge crew were among the closest, I just wish I had known what was going on when I told Benji it would be from 11 to 3. Again, I apologize for all of that. It should have been better handled.

Okay, as for the nasty reply from Neil. Neil was already upset that day because Sci Trek had cancelled on him with such short notice, and to make it all worse for him, I literally chewed him a new you know what when I found out that I had worked all night friday, and the event didn't even go down. I had agitated him to the point of wondering if I was going quit and leave it all to him just about the time poor Benji asked him for information, and before I could shake a cat, we were in utter turmoil, frayed nerves, spent patience, exhaustion set in.

But things are looking much better for the actual contest on Oct 12th. We could use some volunteers to help inform visitors and to help get things going if anyone has some time and can forgive us enough to join us.

I am looking so forward to this, and it is like a dream to me that it is actually happening in Atlanta, after all these years this is the first Loebner Prize Contest that I can actually attend. I almost wish I could compete in it too but that would be a conflict of interests, and besides, many of you have surpassed my original Albert program and this year's contest is a way I can contribute something and try to give back to the contest that has meant so much to me for all these years.

I'd like to congratulate Joy and Benji for their accomplishments and encourage all of you to try and make the Loebner Contest better because I truly believe that these contests inspire people to achieve something that is almost impossible - to seem truly human.

It turns out that there is another female botmaster, named Lynn Hershman, and Robert Wald is her technical advisor, but they put his name on the application.

I like to say that we celebrate humanity with the most sincere form of flattery - imitation, and being human is what it's all about. We share that regardless.

Best Regards,

Robby.

One quickie:

By make programs that try to seem human, we illustrate to ourselves what it really means to be human.

Yeah and being human is complicated

Robby-- Sorry about seeming a little peeved I've only been into bot-building for a half year now, so I had never tried, even with an offline bot, to enter the Loebners; I had no idea how stuff like this 'goes down'
Most of all, don't take our verbal abuse seriously we all just needed a little time to vent... You know
--Shady

Hey Shady,

I understand, it is important to have a place to vent, and to have someone who will listen, and that is the least I could do.

I want to try making a personality forge bot too. I started one tonight but will have to wait till tomorrow to try and buld some language content. I am calling it FRED2002

This is a really neat web site, and I wish I had known about it earlier. I will lay links to it from my web site when I have more energy. I have a list of my favorite AI sites and this is a great one. I like the whole concept.

Robby.

Robby, thanks a lot for coming to the forum to explain things... I think the main problem here was feeling that the contest was a black box and nobody was giving any explanations on anything.

It's true that I have no idea of what sort of emails you've been receiving, but I've seen what people have said in the forums, and I assumed emails would be on the same tone. That is, annoyed but not insulting.

I think part of the problem here is that everybody knew about the Chatterbox Challenge, and that was very well explained. I always assumed that there would be some chatting with the bots for the selection. And after you've seen on the Chatterbox Challenge the scoring system perfectly described, you feel very upset when you find out that nobody explains how your bot is going to be selected.

By the way, will there ever be a list of all the bots that entered the contest, ordered according to their scoring? We'd all love to know how far we are from perfection.

I've just had an idea for next year's contest. Why don't you make a mailing list with all the contestants? That should help to keep communication lines clear.