Simon gets serious with Linear Regression (Machine Learning)

Simon has been working on a very complicated topic for the past couple of days: Linear Regression. In essence, it is the math behind machine learning.

Simon was watching Daniel Shiffman’s tutorials on Linear Regression that form session 3 of his Spring 2017 ITP “Intelligence and Learning” course (ITP stands for Interactive Telecommunications Program and is a graduate programme at NYU’s Tisch School of the Arts).

Daniel Shiffman’s current weekly live streams are also largely devoted to neural networks, so in a way, Simon has been preoccupied with related stuff for weeks now. This time around, however, he decided to make his own versions of Daniel Shiffman’s lectures (a whole Linear Regression playlist), has been busy with in-camera editing, and has written a resume of one of the Linear Regression tutorials (he actually sat there transcribing what Daniel said) in the form of an interactive webpage! This Linear Regression webpage is online at: https://simon-tiger.github.io/linear-regression/ and the Gragient Descent addendum Simon made later is at:  https://simon-tiger.github.io/linear-regression/gradient_descent/interactive/ and https://simon-tiger.github.io/linear-regression/gradient_descent/random/

And here come the videos from Simon’s Liner Regression playlist, the first one being an older video you may have already seen:

Here Simon shows his interactive Linear Regression webpage:

A lecture of Anscombe’s Quartet (something from statistics):

Then comes a lecture on Scatter Plot and Residual Plot, as well as combining Residual Plot with Anscombe’s Quartet, based upon video 3.3 of Intelligence and Learning. Simon made a mistake graphing he residual plot but corrected himself in an addendum (end of the video):

Polynomial Regression:

And finally, Linear Regression with Gradient Descent algorithm and how the learning works. Based upon Daniel Shiffman’s tutorial 3.4 on Intelligence and Learning:

DSC_0557

 

 

Translating Car On Terrain project from Phaser.io into Processing

Today Simon spent hours translating this Car On Terrain project from Phaser.io (where it appears in JavaScript) into Processing (Java). He loved doing it in a form of a lesson for me, while I was filming him and asking questions about loops, arrays, fixtures, shapes and bodies (and there are many things I don’t understand). Simon also spoke about “the three most important properties: density, friction and restitution”. The project involved a lot of Physics, using many Box2D sub-libraries and translating between pixels and mm.

In the end, he got tired of writing all the coordinates for the terrain vertices, but he did get quite far.

 

 

 

 

 

Applying Box2D to translate from pixels into mm:

CarOnTerrain translating from pixels into mm

Translating Bit Invader from Codea into JavaScript

Simon tried to reconstruct Bit Invader game (from Codea.io) in JavaScript, but got stuck at a certain point when he was programming the enemy to recognize the hero and the bullets. Here is how far he got. The project is available on Simon’s page in Codepen:

https://codepen.io/simontiger/project/editor/AdyVmr/

 

Thoughts on Fidget Spinner Simulation

DSC_0420-1
In his recent Live Stream, Daniel Shiffman announced he would make a fidget spinner simulation in the near future. Inspired by this idea, Simon has jotted down some code for the fidget spinner project. The code and part of the code description both come from The Nature of Code, Example 2.6. Simon explains that the code below has a non-static radius, because he is dealing with x and y coordinates and not angles (i.e. not spinning but moving to where the mouse is). He says he is going to use that code as a foundation. The second step will be to rely on angles instead of x and y coordinates. The third step will be to rotate have the fidget spinner keep rotating and to apply a friction force to make it stop eventually.
Simon recorded this video, where he ponders on the possibilities for the fidget spinner code:
Simon wrote the rest of this post:
Description:
This is the code for clicking and dragging an object around the screen.
To do that, you need the following data:
  • mass (radius of the object)
  • location (position of the object)
  • dragging (Is the object being dragged?)
  • rollover (Is the mouse over the object?)
  • dragOffset (offset for when the object is clicked on)

And the following functionality:

  • clicked (checks if the object is hovered over and sets dragging to true)
  • hover (checks if the object is hovered over and sets rollover to true)
  • stopDragging (sets dragging to false)
  • drag (moves the object around the screen according to the mouse)
Code and Pseudocode:
Fidget Spinner Simulation
 – Prepare mouse interaction code
  – Data (mass, location, dragging, rollover, dragOffset)
  – Functionality (clicked, hover, stopDragging, drag)
  – Code (Mover)
   – Variables
float mass;
PVector location;
boolean dragging = false;
boolean rollover = false;
PVector dragOffset;
– Functions
void clicked(int mx, int my) {
  float d = dist(mx,my,location.x,location.y);
  if (d < mass) {
    dragging = true;
    dragOffset.x = location.x-mx;
    dragOffset.y = location.y-my;
  }
}

void hover(int mx, int my) {
  float d = dist(mx,my,location.x,location.y);
  if (d < mass) {
    rollover = true;
  }
  else {
    rollover = false;
  }
}

void stopDragging() {
  dragging = false;
}



void drag() {
  if (dragging) {
    location.x = mouseX + dragOffset.x;
    location.y = mouseY + dragOffset.y;
  }
}
  – Code (Main)
   – Variables
Mover m;
   – Functions
void draw() {
  m.drag();
  m.hover(mouseX,mouseY);
}

void mousePressed() {
  m.clicked(mouseX,mouseY);
}

void mouseReleased() {
  m.stopDragging();
}

Preparing to give Dad some Processing Lessons

Simon has been giving Dad some lessons in Processing (Java), following the chapters in Daniel Shiffman’s book The Nature of Code. The have already finished Chapetr 1 (Vectors) and moved on to Chapetr 2 (Forces). I hadn’t realized until recently however that Simon doesn’t simply download all the examples for the lessons from the Processing website but manually writes them over from the html version of the book instead (to style them the way they are in the book, he said). It’s a amazing how organized he is at doing that and saving his files in separate folders.

Testing his examples Simon actually found a mistake in the book and Daniel Shiffman asked him to create an issue about it on the book’s GitHub’s page: https://github.com/shiffman/The-Nature-of-Code/issues/133

DSC_0120

Box2D into Codea, Java into Lua

Several recent translations of Box2D examples (from Daniel Shiffman’s book The Nature of Code) into Codea, that is Processing (Java) into Lua. Both Box2D and Codea are physics libraries.

Surface:

Revolute Joint (Windmill):

Car:

Autonomous Agents. Steering Behaviors. Flocking.

Simon’s full-screen take on Flocking, based on Chapter 6 from The Nature of Code (Autonomous Agents). The code comes from Daniel Shiffman’s tutorial Combining Behaviors: Flocking.

In Simon’s words, Autonomous Agents need to have action/ selection, steering (for example, for Flocking you need Separation, Alignment, Cohesion programmed using the neighbours, desired and current velocity) and locomotion (position, velocity, acceleration, “own vector stuff” or a physics library).