block by enjalot fd78a64d3ff9f03907c4

polygroups visualization

Full Screen

Prototyping Neural Network algorithm visualization

simulation.json

{
    "polygroup": [{
        "5": [                     // timestep this node was activated
                                  // each tuple describes incoming activations
            [2, 3],                // [node index, timestep when this node fired]
            [1, 4]                // 
        ]
    }, {}, {                    
                                      // this node (index 2) fired at times 2 and 6
        "2": [
            [0, 2],
            [2, 2]
        ],
        "6": [
            [0, 6],
            [1, 4]
        ]
    }],
    "init": [1, 3, 1],
    "numTimeSteps": 6,
    "network": [
        [0, 2, 3],
        [1, 4, 2],
        [1, 3, 1]
    ]
}

NOTES // when a node fires, what would be super cool is in the connection // between two nodes, that node can increase or decrease in size, can // flash different color whether plasticity is increasing or decreasing

Also using @sxywu’s trick of adding links as invisible nodes and centering them between source and target, this keeps the graph from forming straight lines in some configurations.

Using @erkal’s polygon links.


forked from mbostock‘s block: Collision Detection

forked from enjalot‘s block: rigid links

forked from enjalot‘s block: rigid links

forked from mathnathan‘s block: polygroups visualization

index.html

simulation.json