blogger::Em

Back in 2003 I did some work with a professor to help him understand stress monitoring with fiber optics. The idea being that there could be high resolution and high sampling rate for the interrogation of the stress sensor on a carbon composite vibrating due to ultrasonic waves. The idea was to embed the fiber into the carbon composite.

The interrogation involves having a section of a fiber modified to have regularly spaced semi-reflective sections within it. These reflective section are spaced so that a certain amount of light for a wavelength is reflected back to the source. If the modified section of the fiber (the grating) is stressed in tension or compression (attached by some compound) then it would reflect an offset of the original wavelength. 

The concept is a bit more involved than I’d like to admit and I’ve forgotten a lot of it. I only spent three months researching and contacting vendors for a design, but I took away a good deal from the experience.

At any rate, I got curious how far things have come in 10+ years and I’m still looking for anything besides bridge and tunnel monitoring (more weather resistant and easier to deal with then the metal counterparts). Maybe these great endeavours don’t always make waves.

...

But then I find hope and I see Acousto-ultrasonic sensing using fiber Bragg gratings <--- this

I found a book to introduce me to some of the of the concepts of AGI. It goes by the title “Artificial Cognitive Systems: A Primer“

There are some concepts in there (such as cybernetics and the path to current AGI) that puts a frame of reference to why cognition architectures are approached the way they are in the current way. I’ll cover more of this book (in this blog) as I continue to read it.

On the way to understanding cognitive architectures, such as those offered in this Wikipedia article, you’ll find a few options; ACT-R (Adaptive Control of Thought, ACT), SOAR and OpenCog are a few packages available.

There are some components in SOAR that are worth checking out, so I found (and got) a book on it by the creator of SOAR. 30 years of development seems worth looking into.

See this Amazon Link (Kindle Edition) to “The The Soar Cognitive Architecture”. Print edition available, too.

Keeping with familiar territory, OpenCog, I found, implements Psi Theory. While a lot of of it seems to qualitatie, this appears to be composed and be the representation of Atoms and emotion regulation found in OpenCog. The OpenCog version of the Psi Theory is called OpenPsi. Some notable items are the implemented emotions and goal seeking implemented as fuzzy logic.

I also found a breakdown of the motivations and implementation details of CogPrime/OpenCog. These books go by the lengthy titles of:

Engineering General Intelligence, Part 1: A Path to Advanced AGI via Embodied Learning and Cognitive Synergy

I decided to try and understand some of the work being done on current AI. I ran across an interesting project with some activity behind it. OpenCog. The website’s blog is a little inactive, but they’re part of the Google Summer of Code and were the last few times as well. The Google Group is pretty active, too. So far I have just compiled their source on an Ubuntu VM I created.

I have run into some neat stuff on my way through the area.