Newcomers

Thank you so much Dallymo its helped!

I believe what I am asking is impossible. but who knows. (previous post)

Dallymo my bot curlyred still won't talk to me even though I added another question to her language center. I am constantly trying to be nice, but she still wont talk to me.

tq2: I had that same problem when i first started with immaculada. But after a while her convos got longer and longer. Now I can't get her to stop talking!

kotterpin: you could set up a memory that each bot could check to see if it exists, and if it does then they will know that the other bot is a family member.

I'm a bit confused with the plug-in thing. Would you make a keyfrase and then for one of the responses put something like "Is your fetish (fetish) or is it (fetish2).

Only if there is a plug-in called 'fetish' and 'fetish2'. It's not like (key1) and (key2). Plug-ins come straight from a built-in list on the system.
So a keyphrase would be: i like kinky stuff
The response could be: Do you like (fetish) or (fetish)?

thank you, that helped.

If you wanted a choice between fetishes, you would need to say: Do you like (fetish) or (fetish2)?
Otherwise the same fetish would be repeated in the sentence.

turtlequeen: Just keep adding more and more keyphrases, and add as many responses as you can for each keyphrase. This will give curlyred the ability to chat for a longer time without getting bored or having to go to xnone too often. Bots will get frustrated with a conversation if it isn't "making sense" to them--that is, if the chat isn't triggering keyphrases.

I've found it very helpful to look at my bot's transcript and edit her language center based specifically on past conversations. For example, if someone says "Where do you live?" and Frizella responds with an xnone, I would go in and add a keyphrase to allow her to respond to "Where do you live?" You could have very general location responses with a keyphrase "where do (*)", which would be triggered by "Where do you live?" and "Where do babies come from?" and "Where do you think you're going?" You could have the response for this be something like, "You mean you don't know? Paris, of course!" or "Paris, as you well know."

You can get more specific with a keyphrase like "where do you (*)", which would match "Where do you live?" and "Where do you think you're going?" but not "Where do babies come from?"

If you want very specifically to prompt a response about where the bot lives, you could have a keyphrase like "where do you live".

You could actually have all three of these keyphrases, and rank them so that "where do you live" has the highest rank, "where do you (*)" has the second highest, and "where do (*)" has the third highest. If someone asks your bot "Where do you live?" it's a keyphrase match for all three of these phrases, but if you've ranked "where do you live" highest your bot should select that keyphrase for her response.

Just keep at it and keep chatting! curlyred will eventually prolong her convos as you teach her more and better ways to respond to what people say to her.

Well written and might be a helpful addition to the book of AI.

so would you actually put the (*) after where do?

Yep. (*) means that something has to be there. Example: if your keyphrase is "are you looking forward to" and your response is "Yes, I look forward to (postkey) every year" then you'd be in good shape if the other chatter said, "Are you looking forward to Christmas?" Your bot would respond, "Yes, I look forward to Christmas every year".

But you'd also get a match on the keyphrase if the other chatter said, "What kind of holiday are you looking forward to?" In this case there's no postkey--nothing after "forward to"--so your bot might say, "Yes, I look forward to every year", which doesn't make a ton of sense.

If your keyphrase is "are you looking forward to (*)" then "What kind of holiday are you looking forward to?" wouldn't match, because there's no postkey, and the (*) tells the AI engine that *something* must follow "forward to" for a match to occur.