Wednesday, August 23, 2017

The concepts Asa H is learning

I have put a lot of effort into giving (and identifying/naming) concepts that Asa H learns (forms) as a part of its overall "self" concept. This is important as evidence that Asa is conscious. But Asa learns many concepts that are not part of its "self" concept. Some are easy to identify/name. Things like:

bright/light --> hot         ("heating")
black/dark --> cold        ("cooling")
force --> acceleration    (Newton's 2nd law)
day --> night --> day --> etc.     where day=(hot, light) and night=(cold, dark)

It may be difficult to assign human labels (words, names) to some of the other concepts Asa forms.
In some cases these are important; examples of alternate conceptualizations of reality. But certainly not always.

Saturday, August 12, 2017

Reconceptualizing and the game of Go

Google's AlphaGo AI program has developed Go game strategies that humans had not thought up despite 2000-3000 years of collective effort. This is an example of the kind of thing that I have in mind in my work with Asa H aimed at "reconceptualizing reality."

Pain localization

Asa's and the human pain systems are vectoral. For Asa one would like to have one contact sensor wherever (Lego) bricks or (Vex IQ) beams are connected. Temperature sensors can be mounted alongside contact sensors or they can be more sparsely distributed. I.e., "burns" might be a less localized sort of pain as compared with "cuts." Internal pains are literally signaled from internal sensors of either sort.

Tuesday, August 8, 2017

Vex IQ

The Vex IQ brain has some advantages over Lego p bricks. It has 12 I/O ports and is less expensive. One disadvantage is that Vex IQ has fewer commercial sensors available for it. This being said I have managed to hack the Vex IQ bumper switch so that I can use it with my robot pain system. Multiplexers can allow me to connect a large number of pain sensors as needed.

Some effort has been devoted to building robots using integrated Lego and Vex IQ components. I have, for instance, bolted Vex IQ plates (4X_ ) to Lego plates and Lego bricks (1X_). One can then build onto this in both directions using both Lego and Vex IQ parts.

Saturday, August 5, 2017

In times of change

Lots of us have trouble accepting change. But Asa concludes*, logically enough, that you should resist change (offer negative feedback) when things are getting worse and embrace change (offer positive feedback) when things are getting better.

* by associative learning while monitoring utility change

Monday, July 31, 2017

Asa's subconscious

The concepts out of which Asa's self model is made (see my blogs of 21 July 2016 and 1 Jan 2017) and other concepts that Asa is conscious of (see my blog of 19 Oct 2016) are, in turn, composed out of lower level concepts/"features". Each concept (vector) in the Asa H memory hierarchy can evolve over time. The upper levels in the hierarchy are not directly aware of these lower level features and any evolution that is occurring. These details and processes are subconscious.

Saturday, July 29, 2017

It depends on your definition


Whether there is a god or not depends upon your definition of god. (See my blogs of 10 June and 3 July 2017.) Similarly, the question of digital immortality depends upon your definition of immortality. My argument against the possibility of human immortality can also be leveled against AIs (see my blog of  15 Oct 2010). But one might also consider the possibility of an immortal society * of (specialist?) AIs. And if one believes in the block universe interpretation of relativity theory I suppose we are all immortal in some sense even as things stand today. This all points once again to the importance of scientific pluralism.

* perhaps a society that has even formed a concept of its self