@robotmaker I hacked a toy Dalek and added a Sensory (RSC4128) speech recognition chipset it took 2 weeks to make, it was made as a toy concept, but I never managed to sell it.

@PJ another item i noticed on AI software ,to get a really good software since there are a few You need to spend big bucks or get a AI software that fair thats open source.

Edited

@Toymaker yes this is exactly what I have in mind, can easily be interface in EZ-B using the LED sound features to give it that physical realistic presence.

@robotmaker Spending money on big A.I softwares that's not exactly what I am looking for. As you can see in the video I posted in my previous post, I am capable of making my own from scratch to suit my needs. My conflicts with commercial A.I is that they are engineer for the general public use. I am seeking for something more specific. Basically I want SIRI ( apple) or speaktoit ( android) android like feature on my robot using EZ-B and the SDK provided.

The video of the A.I that Toymaker posted is exactly What I am striving to accomplish with EZ-B. I plan on putting together a 5 feet tall robot with built in PC to run ARC , along with a 7 inch display.

Isn't making your own AI the whole idea here?

As you have rightly said PJ, any commercial AI will be very generic and not specific. The opposite to what we hope to achieve. As I have said to quite a lot of people who have asked me if I can give a tutorial on how I've built Jarvis, it's pointless, he is purpose built for my home and my requirements.

Ideas and suggestions may be generic but application will always be bespoke.

Thats the design i am making a 5 foot with a display,plus looking at ways to add emotion to facial on my head.

I know i dont having spending big bucks on a good AI,BUT TO ADAPT them to our design its kinda hard.

PJ did you check the AI link i posted,someone in this forum made it to work with EZB

Now that makes more sense PJ. I did wonder when you put that it is what you are looking at.

As I said, looking at commercial AI and even open source AI is useful to get ideas and inspiration. I know a lot of the personality I will be adding to Melvin is coming partly from AI of NPCs in video games.

Borrowing ideas and adapting them is probably the key to this. After all, we are all borrowing the idea of Jarvis from IronMan at the end of the day.

@Rural, to answer an earlier question. Searching google...

The action is done in Vox Commando. The payload (in this case "charlie sheen") is also broadcast to EventGhost for logging and historic reference. EventGhost passes it to ARC as a variable ($searchterm). ARC can then display this on any LCD or dialogue. It could even react to specific search terms, dance if the search term contained music or if you use the idea I am putting into action on Melvin with mood, different terms could increase or decrease happiness levels etc.

The same action could be done in ARC but the results could be disappointing as mentioned earlier.

Rich ARIEL adapts (by itself) to the primary user over time, so it starts to learn what the primary user likes, ARIEL has even alerted me to science fiction films long before they are released as she watches certain RSS feeds for me as she knows I like science fiction. When a user is first introduced to ARIEL, she asks a number of basic question about the user, from there the user profile is updated as she learns more about what the user likes (ie music tastes etc).

So from starting generic, ARIEL becomes bespoke to any user that she interacts with, the same will be the case with the EZ:1 and EZ:2 robots when the plug-in is completed.

This is exactly how Dragon works it improves the users phonetic profile over time and this means it just gets better at speech recognition accuracy with that particular user, without further training.