Open letter to all contestants
I do not presume to speak for the other contestants, but I harbor a suspician that many of them may have some of the same feelings as I do regarding this contest. I had deep reservations about entering this year, however the need for exposure in a competitive environment overrode good judgment and we did indeed enter. That was in April. It is now September and the transcripts for the initial round are still not published. That is six months. And that is unacceptable. Here is why. Many of the participants including our entry, are commercial projects and we depend on our web presence for sales or for fund raising. How our technology is viewed, favorably or unfavorably affects our ability to successfully operate. It is one thing to place low in the rankings when the transcripts are placed right next to the score. To publish a score that could be interpreted as being a reflection of a poor technology, and then leaving that to stand alone with no explanation on what the questions were, or how they were scored, that is simply not acceptable and a terrible breach of ethics. If the promoters were located in this country, it would be actionable. Reviewing this thread Mr. Lodhi assured me that the initial round transcripts would be published as soon as the initial round was finished but did not follow through on that promise. There have been other irregularities with this contest. For one thing, the the start date is altered, fine but many of us schedule maintenance etc and plan around these dates. When it took forever before running the first round, we asked politely if you could schedule one day for maintenance, but you refused. And so of course the first round managed to take place (at least I am assuming that was the first round) during a scheduled maintenance. Ok, I accept that. At this point, the track record when it comes to entering contests and having ot go smoothly is so bad, I expect it. But, then abandoning the contest with no explanation, just the appearance of a bad showing, unacceptable. There have been questions about this from the last contest, for one thing it appears that the promoters have a relationship with one of the contestants which was never disclosed, which is fine and no doubt innocent, but why the subterfuge? And I am sorry, I know there is no way to lodge a protest without sounding like “sour grapes”, but the questions defintely favored ESL bots and bore no resemblance whatsoever to the way the questions were structured on the samples. In the end, we end up with a really slick looking website with published scores, which stand as a record and which if visited by someone without knowledge of the history, would simply assume was the most legitimate contest ever held, and where the entire thing may have been engineered to give one commercial enterprise the appearance of doing really well when compared with competing technologies. It is higher in the search rankings than longer running contests. So I have no choice but to publicly declare this contest to be a fraud, and I am doing so here first and will also be doing so on our websites and other forums. I intend to disclose that there is a relationship between the promoters and one of the contestants, the fact that the we were questioned during maintenance, after asking for down time, and that these facts are probably reflected in the outcome. I know that in 2012 they finally ended up sending Steve a check, but once again, no transcripts, no record of how the questions were scored, nothing.
In an attempt to correct the record, here is what I am proposing. I would like to resurrect the bragging rights contest for one round. Anyone who is entered in the robochat challenge can be automatically entered, unless you email me, or post here that you do not want to participate. Originally it was a “no money” contest, but I intend to scare up a few bucks, probably no more than $500.00 but no less than $250.00 Same idea as before, everyone provides ten questions\statements, we use a Pseudo random generator to pick one question from each contestant, these questions will be the basis for a single round. As before we will publish a “world view” topic that each bot will be expected to freestyle on, any suggestions for a topic, post them to a new thread I will be starting.will take place just after the Loebner contest is finished, I’m thinkging October 10th. The judging will take place in a single day, and the results will be published and a check in the mail within 24 hours. Thats 24 hours as they occur here on planet earth. The updated URL is
http://contest.risoftdev.com
Vincent L Gilbert
CTO RISOFTDEV inc
RSS AI Zone
