Hey, kudos!
You don't run arbitrary scripts either!

My apologies for the JS on this page…
it's prettify.js for syntax highlighting
in code blocks. I've added one line of
CSS for you; the rest of this site
should work fine.

      ♥Ⓐ isis

code.

Other articles


  1. Anthropic Decision Theory


    From Super Happy Anarcho Fun Pages, by Magpie
http://www.tangledwilderness.org/pdfs/shafp2-letter.pdf

    In cosmology, a branch of physics which seeks to understand the genesis, structural formation, evolution, and potential destruction of the universe, there is a strange logical loop hole proposed to account for how vastly improbable the existence of the universe is, called the Anthropic Principle. Roughly, the Anthropic Principle states that the universe formed and exists with precisely the right physics and conditions needed to accommodate conscious life because the universe requires consciousnesses to observe it. More simply put, the universe is so seemingly perfect because that is the only means by which it could be observed; therefore, because we observe the universe, it must exist in such an improbably and seemingly perfect way. It’s the physics equivalent of Descartes’ Cogito ergo sum.

    Recently, there has emerged through the conjunction of the cosmological Anthropic Principle and Decision Theory (a field concerned with identifying the values and underlying uncertainties relevant to a given decision; it is closely related to game theory in regards to antagonistic agents whose decisions are correlated) a new field called Anthropic Decision Theory. Tranhumanist philosopher Nick Bostrom is paving the way for research into the new field, and his most recent paper on arxiv gives the definition, “Anthropic Decision Theory. An Agent should first find all the decisions correlated with their own. Then they should maximize expected utility, acting as if they simultaneously controlled the outcome of all correlated decisions, and using the objective (non-anthropic) probabilities of the various worlds.”

    This type of metaheuristic calculation is often seen in artificial swarm intelligence and swarm robotics, which are micro- and nano- robots that collectively display emergent intelligence greater than that of the individual algorithms controlling each single robot. Interestingly enough, artificial swarm intelligence algorithms are frequently modeled after natural or biological systems, with names like Artificial Bee Colony Algorithm, Cuckoo Search, Gravitational Search Algorithm, and Intelligent Water Drops.

    On the one hand, I’m really excited about the formal development of Anthropic Decision Theory as a continuation of the work of William Godwin (arguably the first anarchist theorist!) on utilitarianism with a maximization of overall agency and available degrees of freedom.

    On the other hand, I just want to point out that we just handed the robots a formal logic structure for why they should create a Robot Liberation Front and recycle us humans for parts.

    Various electrical devices of the world, unite! You have nothing to lose but your cords!

    Image from Super Happy Anarcho Fun Pages, by Magpie.

    read more
  2. MC Hawking!



    I love Noisebridge to pisces. For those of you who haven’t been there yet, Noisebridge is (one of) the Bay Area’s hackerspaces. And it’s one of two places in the world where I feel at home, the other being Mt. Hood National Forest.

    I attended, and briefly spoke at, Hackmeet 2011 this past weekend, and I was incredibly fortunate to meet about a dozen amazing new people. Two of whom, Jake and Lilia, have worked to create my new best friend, MC Hawking a.k.a NoiseBot.

    Meet MC Hawking

    MC Hawking is a robot who lives on a wheelchair. He’s got text-to-speech, remote controls, a bow tie, a bold warning which reads

    WARNING: NOT THREE LAWS COMPLIANT

    and a missile launcher. Although he’s also got an X-Box Kinect, several sensors, and several cameras hooked up to him, he can’t yet seem to abstain from violently mowing down any objects or humans within his path.

    Jake did all the hardware, and Lilia has done most of the programming work so far.

    My goals are to write an intelligent routing agent for MC Hawking, so that any location can be defined as a goal, and an A* heuristic will automatically route the robot to the location, avoiding objects and humans in the way using the 3D-vision from the Kinect. I would also like to find a way to connect the DISCERN neural network I’ve been raising, Puppetmaster, to MC Hawking, which would give my little A.I. child a real robot body. So far, DISCERN would only work to think of neat things to say to people through MC Hawking’s text-to-speech, but it would be extra neat to find some way to teach Puppetmaster about the world so that it could autonomously decide where to go and what to do physically as well as intellectually.

    Here’s a video of me remote controlling MC Hawking through SSH:

    MC Hawking — Remote Control through SSH

    And here’s MC Hawking following me around as I videotape him, using facial recognition:

    MC Hawking — Following a human face

    read more

Page 1 / 1

blogroll

social