The AI Engine

Kaskroute, that's the channeling bug where the bots get switched sometimes. We are hoping the prof will have time to fix that someday.

Hi, I'm back for who was wondering where I was...

Irina, Prob123, Psim, I'm a bit astonished. You said that soft wildcard are always assigned to a Key? I performed a lot of tests, and always had a different response. If the soft wildcard was assigned, it is in the key list, if not, it was never considered and the key list shift to the left...

I rewrote quite all my bot to avoid this problem!

Why for you it works as fix, and for me no?

There is something that I don't realize.

Hi PROFESSOR, should you explain us how it should work?

Thanks!

Look at my keyphrase. It has a hard and a soft wild card. The soft is always (key1) and the hard is always (key2).

did not you * read (*)

You: didn't you even read the note
Bot: Do you have a copy of the note, I will return it tomorrow.


notice that in one the soft wild card is used It is( key1). the response is
do you have a copy of (key2), I will return it tomorrow.

You: didn't you read the memo
Bot: Sure I read the memorandum,.....honest.

Here (key1) is not used but the memo is still (key2)

Sure I read (key2),..............honest.

Strange Probe, I trust you the most, but I tryed exactly the same and key1 was filled only if a value was given to it. If leaved blank, then, it was not considered at all and key2 was empty while key1 took the value I expected was in key2. My test bot works so, I do not know why!

I'm sure you are right, but to me doesn't work so...

Anyway your comment obliges me to perform some more tests,

Thanks!

Hi All, someone of you will find some strange, horrible and hard transcripts from an user that was in the past assigned to me. Please, do not be angry about that, I friend of mine, expert on software stress test would like to try some of what I think are the best bots.

So he tryied to crash all of them, but he do that from my pc, and so he was assigned to my guest.

Please, use these transcripts only for what they are, just stress test, and about all the bot, in a more or less time, looped... obviously!

Ah, my bot crashed as first (

Can you give an example of your keyphrase to see what is happening?

Prob, I'll give you after christmass. Now I'm dealing with a more urgent problem. Anybody know why if I put in xinitiate a set of KP dealing with crono, when I use to chat, the bot uses them correctly, while if I try to Debug, I had always an xnone?

Message: 'Hello' Time: 35.51
Message: (spell-corrected) 'Hello' Time: 35.52
Message: (preprocessed) 'Hello' Time: 35.53
Name Poss: 3BTESTBOT (Name As Word: )
Begin Sentence Total Time: 35.53
This Sentence: Hello
This Phrase: Hello
Begin Interj Total Time: 35.53
Found: hello+
End Interj Total Time: 35.54
New Sentence: ''
Sentences:
Emotion: 0 Amp: 0 Hello? 1 Goodbye? 0 Yes? 0 No? 0 Haha? 0
ClusterTracker:
Message Info
Emotion: 0 Amp: 0 FoundHello: 1 FoundGoodbye: 0 FoundYes: 0 FoundNo: 0 FoundLaugh: 0
Sentences:
Emote?:

pre-GetKnown: 35.54
end-GetKnown: 35.54
Keyphrase Match:
Blank Message. Ending.

Emotional Analysis:

Favorites?:



What Noun?:


Considered BLAB (1).

Icannot understand why it apply a different rule...

Thanks

Hello is Considered BLAB (1). you can't do a thing it will always go to xnone. I really wish we could get rid of blab and the canned responses.

Yes Prob, but why it consider it blab in Debug mode and work correctly chisung the righg answer when I directly chat with the bot? It is strange...

Hallo, I cannot believe that!
Please, read this debug and help me to avoid that horrible answers!

---DEBUG EXTRACT

This Phrase: "I am sick"
Find: I am sick (9) Time: 37.92
(Found)
Rank & Length Bonus: 9
Position Score: 12 (12 / (0+1))
Sentence Score: 0
(Total Rank: 21)
Highest!

Key: ' I am sick ' PostKeySpan: '-1'
Total Time: 37.92

Emotional Analysis:

Favorites?:


Getting Seek Response
Top Seek: I am sick

AIScript in Responses
Total Time Pre-Chrono: 37.93 RESULTS: '7'
Trying: (injury) (random: 4)

AIScript Memory
ComboScript: ""

Response: (injury)

Before (key)s: 37.93
(ssub): you (1)
(sub): you (1)
(submod): you (1)
(sv): am (2)
(v): am sick (2,3)
(vmod): am sick (2,3)
After (key)s: 37.93
KeySubject: ""
Compound Response!
Random Memory Number: 16 of 28
Memory Chance: 21 GossipTopic: MemoryLevel: 3
No Match. Emotional Response.

AIScript in Responses
Total Time Pre-Chrono: 37.96 RESULTS: '1'
Trying: I love you. (random: 0)

AIScript Memory
ComboScript: ""

Response: (injury) I love you.

----

Given aswer:

You: I am sick, i have gastroenteritis
Bot: in hospital for a month! I love you.

But my sentence was as follow!

+I am sick [0]
I'm really sorry for that!

It is impossible to have sense-full answer if the engine change the meaning as it likes!

Any idea?

BAH!

It looks like you may have the compound sentence box checked on your settings page, that will add an xnone or emote to any short sentence. It works for some bots, I don't care much for it. Just go to settings and uncheck it.
For the keyphrase "i am sick". Try I am * sick with a rank of 30. That covers
I am sick
I am really sick
I am feeling sick.

Okay. I don't know if this question has already been answered, so I'm going to ask anyway.

I know that an OR exists in PF scrips. But do AND commands Exists?

For Example OR is written like so:
<?PF if (mem-example1) is "example1"; if (mem-example2) exists; ?>
The above example would result in giving a response if either were true.

is there a script for AND such as this example:
<?PF if (mem-example1) is "example1" AND if
(mem-example2) exists; ?>
This response (in theory) would only display if BOTH were true and not just one.

So if the AND can already be written please help me. I've been through all the FAQ, AI Book, and most of the forums (from the beggining). Thank You.