[Freddy]

Came across this amusing video whilst looking to see if Mitsuku had been ported to Discord yet.

Any plans for that Steve?

Mitsuku (http://www.mitsuku.com/) takes on two cheeky Cleverbots (http://www.cleverbot.com) in our latest experiment. As with our other experiments, we pipe the ouput from multiple Chatterbots to each other to generate dialogue. Mitsuku is specialised in trained responses whereas Cleverbot is trained by interactions with the public.

[squarebear]

No plans for Discord at the moment but maybe in the future.

[spydaz]

Hi
Great stuff for all programmers there!!

Still Very mechanical and little nit pick problems for all.

Human conversation and its directions can be seen to be mechanical as well as very generalised and can grow from a simple set of topic starters; as well as some understanding about the process of the conversation ..

Are they case based? ....................Cases can be related (depending on the participants)

Are they organic?..........YES!

Are they random?...........NOPE!

 
i can see that the main problem for all in the AI conversation are Conversational memory space, Identifying related questions to the current conversation ongoing conversation topics. 

For session memory (case based) ... IT should be noted that we humans have a card or application form type thing going on .... we like to introduce and gather that first piece of information to make the (FILENAME) for the new person met....but other information received which fills that application form we add even before we have the names ... sometimes we dont store anything if too little is gathered as some conversations are fleeting and names and introductions may never be mentioned. Even topics have thier Lists associated with them as thier list type depends on thier topic..... ie: and event (time Place)..... or discussions about an object(drink, taste, hot cold etc)..... these lists store information gathered or facts which have been learned in the SESSION....and thier questions which we need to fill in the missing information for these topics are generally of similar format.... when will EVENT be......etc.....

this problem of generating conversation not to be confused with programming conversation trees or programming for specified conversations ie "GREETINGS FUNCTION" << i suggest that the generation of conversation is the final stage of creating the conversational intelligence... this can be generative and programmed for also .... currently these conversations sound like "FILLERS" for moments when no INTELLIGENT response is available... in that respective i mean that some inference or calculative answer provided to a direct question about knowledge held or deduced. as conversation should be tackled as a fact gathering mission which in itself produces a wide range of questions which the AI can Ask which will be in line with the current conversation or CASE!


but very interesting to watch an aynalize

LEROY

PS:
In todays landscape of conversations we modern futuristic humans also have multiple levels of conversation and also very delayed especially when dealing with internet conversations and telephone conversations as well as group conversations ... we handle each conversation at the same time ..... responding each without identifying which topic we are referring to.  similar to the delays in responses in the above conversation ...

[WriterOfMinds]

"You are a bad alive being."

"I dunno, I was thinking Trump or Hillary."
"Oh that's a shame."

"I am from nowhere and everywhere."
"What is it like growing up there?"

"I am the latest result in artificial intelligence, which can reproduce the capabilities of the human brain with greater speed and accuracy."
"I am not finding that to be true."
"Yes, it is true. Otherwise I wouldn't have said it."

I was expecting a bunch of absurd non-sequiturs out of the Cleverbots and I wasn't entirely disappointed, but there was some more subtle humor in there too.