Robust-first Computing Wiki

User Tools

Site Tools

Trace:
people:ronnie_garduno

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision		Previous revision
people:ronnie_garduno [2014/08/28 22:06] – created kokopeople:ronnie_garduno [2014/10/06 22:31] (current) – [Project Description] koko
Line 8: Line 8:
  
 I'm interested in AI as a way to understand the nature of intelligence better, as well as a really useful engineering approach. Alife is a really interesting way to further that agenda, to me, in that our current understanding of intelligence seems to indicate that it's an emergent phenomenon, and that would probably be easier to approach from the bottom-up.  I'm interested in AI as a way to understand the nature of intelligence better, as well as a really useful engineering approach. Alife is a really interesting way to further that agenda, to me, in that our current understanding of intelligence seems to indicate that it's an emergent phenomenon, and that would probably be easier to approach from the bottom-up. 
 +
 +===== Project Ideas =====
 +  - **Distributed Attacks:** What are the vulnerabilities of the DHS? The first step to increasing security would be to find these exploitable areas. It might work well to inject malicious data packets into the system, which would try to examine nearby sorters and change their data values to mess up the sorters' performance. It would be best to devise defenses for said attacks as well, naturally.
 +  - **Sensors:** How can we make the MFM interface with the outside world, short of sticking a conventional computer on it and using the output from that? Can we develop a simple sort of light sensor, akin to the simplistic eyespots on some worms or fish? I'd particularly like to be able to setup some sort of reinforcement learning for the system, akin to making a simple plant cell reorient itself to get more light.
 +  - **Diffusion of innovations:** As a totally different idea, I would like to have certain agents in an MFM represent individuals within a society that are trading new ideas. I'd have new ideas injected into the system through some localized agents, representing radios, TVs, or something. Once we could establish a sort of baseline of data on their behavior, we could even study things like the effects of 'censoring' some ideas by making some proportion of the idea emitters stop transmitting them. Should there be varying levels of idea acceptance for each agent? 
 +
 +===== Project Description =====
 +Of the ones above, I've gone with the [[:sensor_array|sensor system]]. Right now the design looks like a phased array of antenna elements surrounded by a grid. The grid is what does the actual sensing, whereas there are emitters of data along the inside of it. Pictures are forthcoming, but anyone can access the source code on GitHub: https://github.com/koko/MFMv2
people/ronnie_garduno.1409263607.txt.gz · Last modified: 2014/08/28 22:06 by koko