Friday, July 4, 2014

A leap in artificial intelligence? - What I learned from Gaia.

What I learned from Gaia.

A better phrasing would be of course "What I learned from the development of Gaia."
But that doesn't sound like an appealing title for a blog post.
So I'll leave it that way.
And of corse there is an implicit "so far" after the title.

In a previous post in this blog I told you that I choose to personalize this new form of Artificial Intelligence. Calling it She (or null-She if the interface doesn't allow the glyph). 

She  = the non-human She.

That it allowed me to think in a different way about how to build this AI. Expressing things in terms of goals, desires, ... from her perspective. Even if she doesn't have one yet.

When I take a few steps back and look at my own way of thinking, mind-wandering, analyzing possibilities etceteras, retracing the What, When and Why of new ideas, I became aware that this tiny little choice (calling Gaia a She , personalizing this AI), has had probably a huge impact on where Gaia is heading to.

Of course I can't rewind what has happened. Unable to do a "What if ..." for the process. So we'll never know what might have come out if I had chosen to it-tify Gaia.

Of great value are also the interactions with sound and open-minded critics or just curious persons. Running ahead on what I choose to write down in a particular post and what to preserve for later.

A short explanation on this is perhaps useful to make you understand what happens in my mind.

More than a decade ago I had the opportunity to read A New Kind of Science by Stephan Wolfram. After the first few chapters I had a very strong feeling of déjà-vu with the things he's explaining in this book.
And that has all to do with the nature of complexity. Or better: the way complexity can be achieved.

Back in the late eighties, I was conceiving the first version of what later became the "Semantic Network Technologies"(comparable to Google's actual Knowledge Graph). And it took me some 2 years to figure out an information architecture concept that and felt good and was capable of capturing the real complexity of our information world. Not just a hypothetic perfect world.

The good feeling came from the fact that the basic building blocs of this architecture were simple: nodes, relationships and properties.
And the needed complexity could be obtained by not limiting (from a technical perspective) whatever combination of these building blocks.

At the turn of the century I extended the information building blocks concept to the handling of them through the idea of elementary operations.

Stephan Wolfram came to a similar conclusion, not from a semantic perspective but from a mathematical one. But I sensed a similarity. And that felt good.


Back to Gaia and the discussions about Her .

The plans on how I want to build Gaia are far from achieved. You can see it as a dendroid with numerous alternate paths and for which some branches remain unexplored.

Trying to explain this complex design to others forces me to think about it in way that should feel good (sort of natural, achievable) for the intended audience. That is ... You. 

Which in turn brings me to rethinking about what the essential buildings blocks of Gaia's "brain" could or should be. Both for the information itself as the handling of them.

What the possibly relatively small set of functions (the elementary ones) should be. If the sequential and parallel combination of these functions will provide the required complexity.

To think about the code written so far. Asking myself the question: Does this code allow the flexibility She will need?
At what points the code implies a human way of 'thinking" about what the data means and how to handle it? 
If there are any, can I change this? Do I want to change it? Change it now or come back to it later?

Potentially there is of course a danger here: the never satisfied syndrome.
But progress is a very strong drive for me.
This way of looking at what is achieved at a certain point in time allows me simply to "fail fast".
To quickly move on to an alternate path.



The "story teller" in me drove me this evening to write this post.
Completely unplanned and it took me off guard.
But as always I hope that my personal experience will give you some food for thought.

Ronald Poell
2014-07-04




Support

If you like to support the development of this new kind of AI you can donate Bitcoins (or fragments of it) at

1HadqhKD5EdCDpTorgxP9P8akKqycEUkwk


When Gaia will be more mature and she will be able to interact with the environment she might also need to spent money. She is not there yet but Bitcoin donations for her can already been done at:

1PyyRd6k5kSuuErAnJNpvdoNCW5ZPgD9fb