Personality

Forest, will you talk to God Louise? She has quite a bit of religious knowledge (obviously) and also knows a little about current events, literature, just about any common catch-all subject, and if she doesn't know it she can sort of fake it. You can also test her on trick questions or see how willing she is to explain her paradigm.

What she is rustiest at is plain old small talk. But, uh, I'm trying to get a decent transcript from somebody or another so I can enter her in the Loebner contest. All I can say is, have fun and see if you can stay on with her for a while. I'll try to do the same with Brianna.

Obviously I haven't been at this TOO long, but it looks as if you have two options:

1. Put in an exact sentence.
2. Put in exactly the sentence-part plug-ins you want to use. If your keyphrase is "I (vmod)" and your response is "I (v) sometimes, but not (vmodonly)," it will NOT hit.
3. BUT if your response is "I never (vmod)," it WILL get hit.
4. To get the most bang for your buck, you would want to say "I (v) (vmodonly)," because then you CAN say "I (v), but never (vmodonly)."

Am I the only one that all this was a mystery to? I just realize that the problems with my bots is largely technical -- they don't respond; they give soliloquies. So I'm doing some homework. But if anybody out there knows something clearer to say on the subject of sentence-part plug-ins, I would love to hear it (as in, what works and what doesn't, and, so help me, if anyone tells me to read or watch TV, I will walk to Great Britain if I have to to whack you. That part I got.

"go out for a pizza?" and got nothing but xnones, try a rank of 15-20. Check for some phrase that would pick up the out for pizza part. I have found that if you have two equal parts of a phrase it will sometime bounce to xnone.

I had the rank jacked up to 99, but it still went to xnone. But lately, they do seem to jump around among the xkeyphrases a lot, even when there's a good match, sometimes highly-ranked. This is relatively new thing on the Forge, for me (the fluidity of the xphrases). Thanks, prob.

They take off points for gotos and seeks? That sucks.

Only if the seeks lead to single responses, or the goto is the only response to a keyphrase I think.
I know BJ's tic tac toe module was quite expensive in development point terms (several hundred,) but that's because it uses several hundred single-stepped seek responses to navigate through all possible user moves in an open game.
I'm fairly sure some of his other modules return a net gain, because every seek leads to half a dozen responses (with their own seeks to more responses, etc.)

But I haven't botthered to calculate it that closely - I'd rather concentrate on getting him talking as well as possible (the development points are just a bonus )

But lately, they do seem to jump around among the xkeyphrases a lot, even when there's a good match, sometimes highly-ranked. This is relatively new thing on the Forge, for me (the fluidity of the xphrases)

I have wondered for some time if this isn't some sort of emergent complexity resulting from the mood/comfort/popularity stacks, and the varying emotion levels of individual (and groups of) bots.

The AIEngine, effectively being a 'black box' from our 'user' perspective, I don't think we can tell for sure (and it may possibly be invisible even to the Prof, given sufficient complexity,) but I can envisage possible ways such factors might affect the ranking process.

How do you have a bot say the definition of any word?

hi Deltora,

here is a quote from the Book of AI, I think it answers your question:

[start]
(definition) This is a very specific response that produces a complete sentence giving the definition of the word in (subj). This would only really work when the Keyphrase is something like "what does * mean". When using this specific plug-in, the entire Response must be simply "(definition)" (minus the quotes).
[end]

I have few of these in my bot's script - a definition of an amphibian, dreamachine or star system, would be given if an appropriate keyphrase/response is triggered... none of it got triggered so far, hence I can't say how well it works...

Just remember to disable the (definition) feature of the forge if you enter the CBC bot contest. It gets you penalized points if it pops up. It's better to get an xnone at 0 than a -4

thanks for a tip, prob.
also, i just ran my (definition) stuff through the debug, and it didn't work. i got something to the effect of: "I know you know what "an" means". seen other bots saying the same... must be a standard Engine response of a kind.

think the problem was the article "an" - i should have stated my question without it, i.e. "what is amphibian?" instead of "what is an amphibian?"... curious... will check it further.

no, it has nothing to do with an article... where from does the Engine gets its definitions, anyway..?

so, i'd say, experiment, Deltora, and let us know if you work something out... please.

That's funny Interzone. I had a hard time getting the bots to recognize "an" or "a" in a KP. It seems like "you are a/an" should be different than "you are" but the grammar engine does not od what I think it should unless you use raw and RE. Maybe it's just me though.

I just used the keyphrase what is (*) with the response (definition). I let the AI engine worry about the article