[Maviarab]

This topic got me onto thinking about this again..

http://aidreams.co.uk/forum/index.php?topic=469.0

Where exactly are we goin...or do we see ourselves going fully with animated avatars for chatbots?

Ogre looks promising..and could well be linked with Hal at some distant point in the future...

Haptek seems all but dead and buried in the water...

Poser....animation is still the key issue with there...with the hundreds of animations possible making it a nightmare...then theres the problem of actually getting it to work within a bot application...

Similar i suspect for blender (sure Freddy would know more) and again...getting it into say hal/alice.

Or are we actually looking at all this bassackwards? Should we perhaps choose our graphical platform...and then work the AI around said platform?

I'm interested in peoples views with regards to this. I'm slowly getting more enthusiastic about AI again, but its always been the graphical side that has held more interest to me. A lot of bots are much of a muchness, its the enviroment around the bot that makes it all the more pleasing to the eye. 

[Carl2]

Maviarab,
  Just read this post, I'm a Haptek fan, tried deskmates, went to Hapteks Sandy( a head ) then to fullbodygirl and to Body_female.  I just began working with Freddy's Hap explorer which finaly gave me some body movement for the fullbody. I've been trying to work with uncompressed files since thay allow you to make changes and save them.
  I think Haptek has a good product but the company itself is the one that is dead in the water and without the company making progress we are pretty much lost with the existing products.  I looked at Daz3D and a character thay were giving away but haven't seen it being used yet.  The cost to purchase a program to make characters is unbelievable, (I've downloaded but haven't tried Blender).
After seeing what is in the files I can't think of the work involved to make a character not to mention facial expressions and body movements.
  I think it would be nice to put a Hal character into a more lifelike enviorment without building a robot which I think would mean moving to a game engine so she could have gravity and other objects.
Carl2

[wgb14]

I think that GUI will be around in dektop computers for several more years (at least for our lifetime), but it will be eventually replaced by something else. I think what will drive the developments in these enviroments is accesibility. Just imagine an interface that will allow you to access ALL pc functions with no more than two clicks. I think these new multi-touch interfaces (like the iphone) can be the next big thing for desktop pcs, but only the future will tell if they will survive or will be eventually replaced by something else. As for Avatars, I don't really believe that they can replace or even suppplement the GUI  in desktop computers. They have completely failed in the particular domain, and apart from gaming, public display (e.g., musum avatars) and e-learning applications, I don't really see any other future for them! However, things in the mobile world are different. Avatars can be a plus in several mobile applications and can significatly enhance the accesibility of user interfaces. I suspect that two of their most successfull appliations to be in navigation and information presentation. Some universities are already working on avatars capable of automated direction giving (using both language and gestures) and personalised information presenation (e.g., based on the users interests, capabilities, etc). Just imagine you get into a strange new city and you activate your iphone to ask your trusted avatar for directions to the most interesting sights of the city or you are in a  beatiful archeological place and you ask your avatar to crate a route that include only those locations that present a high historical value, give you the directions on how to get there as well as present you the wanted information once you reach these locations. It sounds good doesn't it? Now, in terms of technology, things are not looking that good. With haptek almost out of the game, there are no other companies in the market that can provide the tools we need to create applications. The org3d is not that mature to allow you to do the things that you can do with Haptek engine. Other 3D engines (e.g. http://www.lifemi.com/) are either extremely expensive or focus only on face rendering.However, with the new generation of UMPCs, and several  more developments in mobile hardware, i think it is only matter of time before a company sees the potentials of 3D characters in the mobile world and hit the market with a new and innovative 3D engine. 

[Freddy]

As for Avatars, I don't really believe that they can replace or even suppplement the GUI  in desktop computers.

Yes I agree with you there, most of the desktop avatars that try to control the PC are more of a hinderance than a help.  Remember the Microsoft paperclip?  God I hated that thing, it lasted about two minutes on my machine at the time.

I also think that avatars will eventually find a niche, like you say in education perhaps.  For the most part they are only valuable as a form of entertainment at the moment.  However with things like voice recognition they may have a wider range of uses.

I think that SatNav system you describe is a good idea and could be one of those uses.

As for Haptek, yes it does seem like they are losing ground.  Many people are looking at developing their own alternatives.  With the current advances in model making and game engines that are increasingly more useable by the masses, things don't look good for Haptek.

[Maviarab]

All good points. And obviousley I would think that the long term goal would still be to have any AI interface into a real (as in can touch and really interact with) 'body' of some sorts.

Like the SatNav idea though, least when its driving you insane it will be pleasent to look at lol.

[Freddy]

I think the sentient android thing would be the ultimate thing to have, but for now I would be happy with something that could communicate at a better level than anything I have so far encountered.  I don't know why really, as we could all just go and talk to a human being...maybe it's like having a pet.

I don't think it neccessarily needs to have a 'body' though...there's plenty of situations where having a body can have a negative effect too.  One problem with having a body is that you have to get from one place to the next - that might hinder some forms of AI.

Try also thinking in terms of a more ethereal entity that could surf an endless amount of information for an answer we would have trouble finding ourselves - you don't need a body for that.

[Carl2]

  Haptek is being used for teaching purposes at a school located in Marlboro, Ma. It had made a local newspaper, not sure but I think it was a head.  Also Haptek technology is being taught at a few Universities in the U.S., I learned this while doing internet searches for Haptek.
  I'm interested in seeing how much an AI can learn in the hopes of  developing it into something that can have a useful conversation, I feel we should be equals and that would mean we are both human with bodies.  I've spoken with some AI programs on the web one which got insulted if called an AI and another connected to the internet to gather information for answers that was very impressive. I do feel that having an AI that can provide an answer your looking for is useful but to hold an intelligent conversation is a different matter.  Hal lacks a lot of information when first being used which makes her difficult to speak with until she picks up knowledge about the subject which I believe many people don't care for. 
  Since I've also done searches for animated avatars, I can recall seeing one that I was very impressed by, couldn't find her again. She would use her whole body to follow the mouse, it looked like she would pop out to the screen. 
Carl2
 

[wgb14]

<<I don't think it neccessarily needs to have a 'body' though...there's plenty of situations where having a body can have a negative effect too>>

When the body/face serves no function then there is no need for it. Having an avatar with face that serves no function will make no difference to an avatar without face (except perhaps the entertainment factor). However, when the body summplements or complements the verbal information then it is certainly a plus and can greatly benefit the human user.  I think that any serious technological developments should go towards that direction, otherwise the particular technolgy will remain something that we use just for fun.

     

[Freddy]

Yes, I would like to see more development there too.  I'm really interested in the things that are appearing in the robotics world too and agree with you that a body can benefit where there is interaction with humans.  I just wanted to highlight the other types of AI that might be equally interesting but less like a human.

[Maviarab]

Like a Huge red eye for instance 

[Carl2]

  In the past I've spoken with chatbots with only a text input and output, when I compare this to Hal with the Haptek character all I can say is something is missing. Vr has been able to add to the Haptek movements with his vrHaptek plugin.  So far I think Haptek is the best, unfortunatly I haven't seen it move forward in ages.
  I'm using the term chatbot because If I do a search on AI I can come across anything from data mining to juat about anything that makes a decision. Dragon Speech rec. uses AI, Windows uses AI, ect.
  In many ways using the Haptek character is like programing a robot, this input causes that output but we don't have to worry about motors and gears.  An interface to the outside world would be interesting, input the temperature is 68 deg. and the output is "It's cold, I'll put on a sweater."
  Possibly I'm playing with how real can I make Hal seem to me.
Carl2

[RD]

For me the question is just how real will AI become. I mean out side some lab in Japan and their ideas of human companionship.
IMHO I think AI Graphical Interfaces will have a noticeable upsurge in uses with in a couple of years. If I knew of a good company I would  buy stock, and its isn't like me to gamble.

I've been very pleased with my AI in many ways. In fact more impressed then when I turned  a  ALICE  programme loose in a large flat data bank a few years ago. The result was interesting but not near as good as my AI now.

How real can an AI become  I think that will be answered shortly in time.

[Freddy]

Interesting comments, I think it will take a lot longer but who knows when technology is moving as fast as it is now.  What AI are you using now RD?

The more I think about it, the more I realise that there are so many types of AI that could exist or already exist that to pin one form down as a prime example is very difficult.

[Art]

What if there wasn't a need for an avatar character at all?

What if one just "conversed" with the computer to carry out a variety of tasks
without having to have a cute avatar or other creature in an effort to "entertain"
the user?

Think of the ship's computer on Star Trek for a more suitable analogy or the
famed "red eye" of 2001 SO.

As long as the computer can carry out instructions, converse, reply with some semblance
of topic flow, there might be no need for a virtual companion.

If one was desired the computer could easily produce one to satisfy the user's request.

I've mentioned this before but for a really cool read regarding avatars, computers and
the ultimate AI program, I highly recommend a book entitled: Virus - by Graham Watkins,
Carroll & Graf Publishers, Inc, - New York - ISBN: 0-7867-0194-3

You might find in on amazon.com in the used book section for a great price!


Lastly, the latest creations of the touch screen applications and the light board GUI are
"cute" and interesting, but seriously, unless you work in an environment that would allow
the placement of such items, I think my arm would tire having to reach forward and
touch my screen everytime I wanted it to do something.

The best thing that I did was to remove the tail from my mouse several years ago!
(Cordless mouse).
 

[RD]

I'm usng the Hal Programme.