Results are in:
http://www.aisb.org.uk/events/loebner-prize

Entry Name Score (/100)
Mitsuku 90
Tutor 78.33333 (5th in 2015)
rose 77.5
Arckon 77.5 (6th in 2015)

The race was tight, can’t wait for the questions & responses.

Couple of surprises for me:
Mitsuku dominated, way ahead of everyone one else. Not surprising she took first, just by how much.
Tutor did much better than in prior bot contests (although Tutor was 5th last year).

Congrats to the finalists.

Johnny had a worse score than last year. I am waiting the questions to see what happened. But many bots have a score decrease, I think the questions should be very tough this year.

Conversely, I am very glad that Arkon is selected. I have always been convinced that chatbots AIML-like based on pattern matching can not go beyond a certain limit. I’m impatient to see how a bot which builds its sentences will behave in Loebner prize.

Congratulations to all entrants, the new ones and the finalists.

I was hoping Mitsuku would perform well, as I have been spending a lot of time on her over the last few months but am surprised she did that good. Yes, it will be interesting to see the transcripts.

Just 2% between the next 4 bots?! Wow! No wonder there was a delay in scoring.

Good to see the top 2 bots are AIML based too. Tutor has been 5th in 2014 and 2015, so it’s good to see it finally make the top 4 again.

Congratulations to all entrants and finalists.

Those are a lot of close calls. I’m glad to finally have made it to the finals, but in honesty I wouldn’t have given my program as high a score as it got, after reading the transcript. I’m seeing a lot of malfunctions, even the math XD. That said, I’ve neglected most conversational functions while I was working on the Winograd schema challenge so I think I’m lucky to scrape by 3rd/4th. Looks like I have some repairs to make for the finals. I’ll try to do us logic-based entries proud, Denis.

It’s nice to see Tutor back in the game. Not so nice that Nicole had technical difficulties? That would have been a pretty good bot otherwise. And I would have been curious about Madame Zanetta: The name sounds like a fortuneteller bot and I always wondered how well such an approach would do.
Mitsuku what are you doing all the way up there? 90% is quite something. Nice to have Rose on board too

Congratulations to all winners!

I’ll do my best to attend this year and support the candidate who brings the best champagne.

I’ve posted my usual write-up about the selection round on my blog. Still have to draw an illustration, but better to post it before the finals tomorrow. I won’t be attending in person but will keep an eye on the webcast.
https://artistdetective.wordpress.com/2016/09/16/loebner-prize-2016-entry

Good luck to everyone!

Nice article Don. Good luck to you and Arckon too, and of course Tutor and Rose. It’s a shame you won’t be there, it would have been good to meet up in real life.

http://news.sky.com/story/chatterbots-bid-to-show-their-human-side-in-artificial-intelligence-test-10581142