Skip to content

Shark-y/VoodooNet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

1 Commit
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

🎩 VoodooNet: The Magic Hat of Neural Networks

VoodooNet is a non-iterative, zero-train neural architecture that replaces the "thermodynamic cooling" of Gradient Descent with instantaneous Galactic Expansion.

By projecting data into high-dimensional manifolds and solving for weights using the Moore-Penrose pseudoinverse, VoodooNet achieves superior accuracy to standard SGD in a fraction of the time.

πŸš€ Key Results

Dataset Method Hidden ($d$) Accuracy Training Time
MNIST VoodooNet 4,000 98.10% < 1s
Fashion-MNIST VoodooNet 4,000 86.63% ~94s
Fashion-MNIST SGD (10 Epochs) 64 84.41% ~9s

Note: VoodooNet hits "Hyper-Galactic" accuracy on Fashion-MNIST without a single step of backpropagation.

🌌 How it Works: Galactic Expansion

Unlike traditional networks that "learn" features slowly, VoodooNet treats the hidden layer as a Magic Hat.

  1. The Expansion: Inputs are projected into a massive hidden space ($d=500$ to $4000$) using high-entropy random weights.
  2. The Discovery: We skip training. Instead, we use a closed-form analytic solution (Pseudoinverse) to discover the output weights in a single matrix operation.
  3. The Scaling: Accuracy scales near-logarithmically with dimensionality ($Accuracy \propto \log(d)$).

πŸ› οΈ Installation & Usage

git clone https://github.com
cd VoodooNet
pip install -r requirements.txt
python run_galactic_mnist.py --dims 2000 --dataset fashion

About

Achieving Analytic Ground States via High-Dimensional Random Projections

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages