Chip Vivant By Mohan Embar wins Loebner prize
1st Mohan Embar 1.25 points.
2nd Bruce Wilcox 2.
3rd Daniel Burke 2.75.
4th [url=]Marshall Allan 4 (no profile yet)

Thanks to everyone. I’m beside myself with joy. This year’s judges were absolutely wonderful. They knew Chip was a chatbot, but were understanding, indulgent, and looked for the right stuff.

I badly need to update chipvivant.com with my 2011 experiences and other stuff I’ve written since then.

I’d also like to thank Bruce Wilcox for his amazing work. This year, I had to choose between shoving lots of canned responses into Chip or working on the real AI engine more, and I opted for canned responses and shamelessly appropriated Bruce’s conversation topics and concept lists. (Chip Vivant doesn’t use ChatScript or Bruce’s engine though.) It’s a true sign of his magnanimity that he gives away all the stuff he does for free.

Also, thanks to Hugh for doing this year in and year out despite the naysayers.

Also Erwin, I know I haven’t posted enough, but does this win qualify for your removing the restriction on my specifying a link in my profile ?

People - I’m still in shock with what just happened here. One of my main criticisms of this contest is that despite Hugh’s vision, the judges’ slavish interpretation of the contest rules favored fake backstories, canned responses and other trickery over real effort.

Well, Chip won:

- without spelling mistakes or fake backspaces to correct fake errors
- despite saying stuff like “I didn’t understand what you just said” and “I can’t deal with that syntactic variant yet - instead of “Jim likes peaches?”, use “Does Jim like peaches?”
- despite his inability to say what his profession is (let alone, mother, father, brother, dog’s name, sister-in-law)

This flies in the face of a lot of long-cherished beliefs people have about this contest, including my own. I don’t know if it’s a fluke, but all of the judges were pretty consistent in how they approached this.

—Mohan

Congratulations again Mohan!

Unlike the previous year I attended the contest (in 2009), I now really understood what was going on.

I personally stood behind one of the judges who said: ‘hey, you’re actually seeing the typos I make!, different from the well known Instant Messaging clients who only pass input after an carriage return has been given.

But I still believe the The Loebner Prize Protocol (LPP) is ahead of IM. Just like in a normal conversation, every word, or parts of words can’t taken back; one pronounced, it’s being heard by the receiver. So I think that’s a pretty cool part of the protocol.

Nowadays, it can be used however to distinguish a chatbot from a human participant, but as you said, the judges were fair enough to look at the best quality. No chatbot in the world is yet capable of fooling a human judge, and you simply did the best job possible! So congrats again!

Erwin Van Lun - May 16, 2012:

Congratulations again Mohan!

Thank you!

Erwin Van Lun - May 16, 2012:

But I still believe the The Loebner Prize Protocol (LPP) is ahead of IM. Just like in a normal conversation, every word, or parts of words can’t taken back; one pronounced, it’s being heard by the receiver. So I think that’s a pretty cool part of the protocol.

...

Nowadays, it can be used however to distinguish a chatbot from a human participant, but as you said, the judges were fair enough to look at the best quality. No chatbot in the world is yet capable of fooling a human judge….

That’s precisely the point. I agree that the LPP is capable of conveying more information than an IM-based protocol, but I also believe that given my goals, it would be an unnecessary waste of my time to devote energy to fake backspacing over fake errors when it’s ridiculously easy to spot the bot in other ways. Baby steps.

congrats Mohan, nice job.

Jan Bogaerts - May 16, 2012:

congrats Mohan, nice job.

Thanks, Jan!

Some interesting reading.

http://www.newscientist.com/blogs/onepercent/2012/05/chatbots-still-fail-to-convinc.html

http://people.exeter.ac.uk/km314/loebner/index.php

Many congratulations Mohan,

From the logs I saw, it definitely appeared the most human of the 4 entries and deserved to win. One bot (Linguo) merely spouted unrelated questions back to the judge regardless of what he typed in and I’m not sure if Bruce sent in the correct version of his bot, as it was trying to convince the judges it was a cat rather than a human.

An interesting day though. That’s for sure.

Congratulations Mohan,
PD: your beliefs are close to mine and the judges did it well this time, it’s better to build a stupid chatbot who at least ‘knows’ it and is humble, rather than one that keeps guessing and gives only elegant or outstanding answers..

Which Chat engine does Chip Vivant use? Mohan clarified it doesn’t use ChatScript. Then which one?

I believe he built his own engine from the ground up.