dandavison

Bayesian Frameworks in Cognitive Science and Computer Vision

A conversation about the free energy principle, predictive processing, and how these theoretical frameworks compare to practical engineering approaches in robotics and autonomous vehicles.

---

Q: Are you familiar with explanations and theories in cognitive science involving "free energy"?

Yes, I'm quite familiar with this area. The free energy principle (FEP) is a theoretical framework developed primarily by Karl Friston, a neuroscientist at UCL, that attempts to provide a unified account of brain function, perception, action, and learning.

dandavison

<script type="text/x-template" id="vue-file-agent-demo">
  <div>
    <VueFileAgent
      :uploadUrl="'https://www.mocky.io/v2/5d4fb20b3000005c111099e3'"
      :uploadHeaders="{}"
      :multiple="true"
      :deletable="true"
      :editable="true"
      :meta="true"
      :accept="'image/*,video/*,.pdf,.zip'"

dandavison

fn prepare(line: &str) -> String {
    if !line.is_empty() {
        let mut line = line.chars();
        let prefix = line.next().unwrap();
        let output_line = line.collect::<String>();
        format!("{}{}", prefix, output_line)
    } else {
        "<empty line>".to_string()
    }
}

dandavison

dandavison

#!/usr/bin/env python

from scipy.optimize import bisect


def record_optimization(func):

    def recorder(self, *args, **kwargs):
        if not hasattr(self, '_recorder_inputs'):
            self._recorder_inputs = []

dandavison

#!/bin/bash
set -e

# DOWNLOAD SOURCES FOR LIBEVENT AND MAKE AND INSTALL
curl -OL https://github.com/downloads/libevent/libevent/libevent-2.0.21-stable.tar.gz
tar -xvzf libevent-2.0.21-stable.tar.gz
cd libevent-2.0.21-stable
./configure --prefix=/usr/local
make
sudo make install

dandavison

dandavison

Two columns satisfy a "one-to-one" constraint if each value in column 1 is associated with exactly one value in column 2, and vice versa. E.g. in this data, the person column and the email column satisfy a one-to-one constraint:

>>> table = [('pablo', 'ppicasso@gmail.com'),
             ('georgia', 'gokeefe@hotmail.com'),
    	     ('mark', 'mrothko@aol.com')]
>>> is_one_to_one(table)

dandavison

width = 500
height = 300

cube =
    nodes: [
        {x: 0, y: 0, z:0},
        {x: 0, y: 0, z:1},
        {x: 0, y: 1, z:0},
        {x: 0, y: 1, z:1},
        {x: 1, y: 0, z:0},

dandavison

class BaseChart

    constructor: ->
        @_width = 720
        @_height = 80

    draw: (selection) =>
        me = @
        selection.each( (d, i) -> me._draw(this, d, i) )