Feature suggestion (sort of...)

[VincentReaper]

Hello,

first of all sorry for that pretty vague topic title, I couldn't wrapp my head around a better and more apropriate one.



Recently I stumbled upon a pretty interesting project called 'ejTalk' (spoken: edge talk), which got by far not enough attention in my opinion, even though it is a very promising project and far ahed of its time.
As I watched a video where Cassandra (that's the name of the flgship project) has been tailored into a virtual assistant, I eventually had to think about Braina and its possibilitys and potential.
I really recommend watching the video: https://www.youtube.com/watch?v=7SWGQHWetRI
(Aswell I recommend watching the other short videos aswell: https://www.youtube.com/user/industrialpoet/videos?sort=dd&shelf_id=0&view=0 )

A snippet from one of the FAQs sections:

What does ejTalk do?

We do several things but they all center on human-computer interaction as a natural conversation.
Over the years we have created and continually enhance a Conversation Management Engine which we call ejTalker. It is a standalone, compact executable written in C++. The ejTalker language defines how a conversational interaction unfolds. It coordinates the modalities (speech I/O, GUI, Sensors, Avatar, external data sources, etc.) and remembers the interaction with short and long term memory. ejTalker manages the thread of the conversation. It is the intelligence behind the speech and graphical presentation of an interaction.
We consult on natural and multimodal systems for established products/systems as well as for new projects and startups. We can help with any system interaction that will benefit from reduced cognitive load.
While we don't create or provide the I/O technology that ejTalker manages we are very familiar with these technologies. So we can provide rich and deep consultation services on these indivual technologies too.

What technology is ejTalker based on?

ejTalker is driven by its own declaritive specification language. The formalism is XML and so is very compatible with all web contexts as well as embedded contexts too. ejTalker does not use a single concept. It is not just a rule/case/neural-net/Markov/cfg-grammar/stocastic/etc. system. It uses all of these and more in layers and combinations. By themselves the individual technologies have strenths and weakness.

(Link to the (probably by now outdated) ejTalk page: http://www.ejtalk.com/v2/home.html )



What really got me hooked was the great conversional capabilities of the project that lifts speech recognition and handling with the AI to a completely new level not seen anywhere else right now (not even with Braina to be honest).
The talking to the program becomes MUCH more natural and conversational and the responses and personality of the program are plain outstanding by todays standards, even though the technology used is nothing new nor special.
A second really interesting part is the chatting ability of said program that just flows realy naturally and conversational, aswell as the ability to link a sort of emotions and personality to the output that is further enhanced by the simple but powerfull avatar used. (On a side-note: researches and studies showed that human beings tend to relate much more to AI programs using some sort of visual avatars. But that topic would go too deep into psychologie and would went too far off-topic.)

Well, I could go on for quite a bit more but that's not the point.
I would suggest watching the short videos and get some inspiration off of it, or maybe even get into contact with Emmett Coin (the creator of the project) and work toward a colaboration, as these two projects together (Braina and ejTalk) would not only work out perfectly together but also would probably get the whole art of AI quite some levels further ahead

With kind regards
Vincent



On a sidenote:
As I write all this I get to wonder if it's possible to customise Braina more than just with the in-built customisation dialogues like custom commands etc. ?
I start feeling the urge to try and tailor some sort of visual avatar plus some sort of intelligent chatting ability onto my Braina Pro program
Any possibilitys and/or recammendations for that?

[saurav]

@VincentReaper  Thanks for the post. We can't find a download link for the program. The program doesn't seem to exist in the public domain.

[VincentReaper]

Sadly that may be true, at least I can't find an official downloas aswell.

What I guess from going through the videos (so I'm talking from a point where I just read between the lines) is that he's trying to find soeone (either person, company or project) to colaborate with, before giving out his code publicly.
But I could be wrong on that.

I just thought I'd put this up as a suggestion, because of the possibilitys that would arise from the combined potential of the two projects. Or as a sort of inspiration to draw from.