Embedded Files
Fibanocci with no pause = double.
This class reproduces the diagram above.
A numbering system to identify any member and to advance cycles.
Placement Physics
Placement Physics
Working on 3D physics to evenly spread out nodes using fixed links that rotate and any angle from the source. Avoidance means just getting a minimum distance away at the opposite angle from any nodes within the minimum distance, while maintaining the link length (radius).
Fibonacci style tree with names. Blue are adults, reproduce each cycle. Newborn (white) add zero, don't reproduce next cycle.
Interesting design for tent with full slit opening, totally procedural, using Fibonacci series.
First load of data in bottom layer, first half true (green), rest false (red). Nodes activate with any true input (for first run).
3 false 1 true data.
Same data with phase shift of 3, seems to retain information.
Wrapping tree radially. Functionality is the same.
Fibonacci style layout
Fibonacci style layout
- Activation: any true passes
- Data is boolean
- Each layer (level) count is a fibonacci number (1,2,3,5,8,13,21,etc.)
Addition
Addition
- Sum Activation
- Studying different level values
Streaming Data:
Streaming Data:
- Cosine wave as input
- Noticing the different cycles
Check out how the bottom value kind of stabilizes.
Wave patterns distributed to result.
Can this be used as a neural network?
New Focus:
New Focus:
- Practice illustrating nodes and node interactions
- Show data travel thru nodes graphically
Minimum and Maximum Distances:
Minimum and Maximum Distances:
- smoothing between two extremes
Ready to simulate orbits.
- number of balls is 4 pi rad squared ( x 4 )
- radius of balls is 1 / (distance squared)
History Approach:
History Approach:
- in 2D, transmitting vectors
- each cell is a transmitter and a receiver
- each history level is one unit distance
Linked:
Linked:
- Checking neighbor only to transmit energy,
- Perfect alternative to quad and oct trees.
- Runs 25x25 at 5fps
- field approach vs point approach
Spheres.
- Field members act like neuron weights with memory.
- Field retains traces of historical data.
In 3D
Page updated
Report abuse