You think so? Did a pure AIML bot not recently win the Loebner Prize for the first time in years?

While it’s true that a pure AIML chatbot won the Loebner last year, it’s also true that the chatbot that did the winning is undeniably the best of it’s kind, and reflects a level of mastery, craftsmanship and ingenuity not seen in any other AIML chatbot. And no, I’m not sucking up to Steve. I would normally avoid such accolades, just to make sure he didn’t go getting a big head, but I felt it to be more important to state the obvious and risk the ego boost than to pass up the opportunity to simultaneously pat Steve on the back and tweak his nose.

Rare would be the day that I would be impressed by a computer’s ability to calculate

I’m impressed with the outstanding scores on the list though. I regret that Izar was disqualified for not being available at one particular time. Combined with the others’ expressed needs to take their chatbots offline temporarily, I see room for improvement in the organisation there.

Carl B - Jun 15, 2014:

∞Pla•Net - Jun 14, 2014:

When your bot can do integrals, or even “solve for x” type of queries I’ll be impressed

SAIL Should be able to handle that. As mentioned in other posts, we had some problems in the middle of the contest which demanded some down time, but we should be back online by the 23rd of this month. If you are interested, you can prepare some math queries and we will see how she handles them. For really complex equations you need to play with the intermediate language syntax a little bit to get used to it http://ai.r-i-software.com/TypeVIe/math.asp but its not too hard. Multiline is OK, should recognize the basic symbols for PI etc…. other notation such as Rad, Tan etc…

Vince

Carl B - Jun 15, 2014:

Maybe so, but pure AIML bots are rapidly becoming less competative.

Interesting. As well as winning the Loebner Prize, the AIML bot in question has also won the Funniest Computer Ever contest for both the years it has been running and is also currently defending its crown in the RoboChat Challenge.

While I of course appreciate the ego massage from Dave, it’s not just my AIML bot that does great in competitions. 3 out of the 4 finalists in last year’s Loebner Prize were AIML bots. AIML is the only technology to have won the Loebner Prize on more than one occasion by 2 different botmasters.

As I have always said, blame the artist not the palette.

@Dave RE: Steve Nothing wrong with graciously acknowledging someone else s hard work and obvious success. Besides, considering Mitsukus track record when it comes to winning cash prizes, if we all did manage to meet at Bletchely Park or some other venue, he may very well be the only one who can afford to buy a round of beers

Having attempted to organize a contest, put in considerable work only to have 2 entries enroll, I can appreciate how little prompters appreciate a “whiney diva” , so again, not criticizing the current contest or scoring, but when you review the comments on contests in general, including the Loebner prize and the proposed x-games prize, it seems that there is a general consensus of opinion that a better format is needed, and a decided lack of consensus as to what that better format might look like.

VInce

Most good bots are a reflection of the size of the knowledge base and commitment of the botmaster. AIML is verbose, but it can do basic math. If the interpreter also supports the JavaScript extension then the sky is the limit.

The ability to handle natural language math statements is a non-trivial task. I started a NLP math thread to help others explore this topic..
http://www.chatbots.org/ai_zone/viewthread/1731/

Steve Worswick - Jun 15, 2014:

Carl B - Jun 15, 2014:

Maybe so, but pure AIML bots are rapidly becoming less competative.

As I have always said, blame the artist not the palette.

Yeah Steve, I think you have it right there- my mistake

Carl B - Jun 15, 2014:

  When your bot can do integrals, or even “solve for x” type of queries I’ll be impressed

I’m impressed by your contribution to this thread right now, Carl B.  That gave me an inspiration. In theory, it may be possible to teach a chatbot about at least some of the basics of Algebra.

By the way, you are making me laugh when I read your comments on this thread.

∞Pla•Net - Jun 16, 2014:

Carl B - Jun 15, 2014:

  When your bot can do integrals, or even “solve for x” type of queries I’ll be impressed

...That gave me an inspiration. In theory, it may be possible to teach a chatbot about at least some of the basics of Algebra.  .

Glad you are staying inspired! That is what these contests are usually good for- getting inspiration.

Btw, since Merlin recently started a solid thread regarding NLP and math functions, probably should post any more on that in his thread (if u already hav’nt). .

Something basic like this simple example, should work.
I just have to figure out the NLP for the chatbot interface.

Mr Lodhi,

Still waiting for those last 2 contestants. Fair warning, I’m going to give it a few more days before publishing my transcript and an analysis of SAIL’s performance.  These are commercial ventures Mr Lodhi, and again I am not displeased with the performance in light of an informed analysis of the transcript, but just leaving in the public view as a low score with no explanation or transcript is something that I cannot do.  So 2 more days and I publish.

Vincent L Gilbert
CTO RISOFTDEV inc

Vincent Gilbert - Jun 17, 2014:

Mr Lodhi,

Still waiting for those last 2 contestants. Fair warning, I’m going to give it a few more days before publishing my transcript and an analysis of SAIL’s performance.  These are commercial ventures Mr Lodhi, and again I am not displeased with the performance in light of an informed analysis of the transcript, but just leaving in the public view as a low score with no explanation or transcript is something that I cannot do.  So 2 more days and I publish.

Vincent L Gilbert
CTO RISOFTDEV inc

Dear Mr. Gilbert,

We would like to inform you that we will publish the transcripts, as soon as the last two bots are evaluated, so please be patient, as it would be highly appreciated.

Kind Regards,

Robo Chat Challenge Team

Dear All,

Elementary Stage is done & dusted and results are up for all the bots. We would update the transcripts on our website once we are done with Final Stage of Robo Chat Challenge 2014. For more details, please visit: http://robochatchallenge.com/

BR,

Robo Chat Challenge Team

Dear RCC,

Please recalculate the score for Ashly
to see if you can at least increase the
score for Ashly to 211 or more.

Elizabot wants last place instead.

8pla