Skip to content

Latest commit

 

History

History
36 lines (25 loc) · 1.46 KB

File metadata and controls

36 lines (25 loc) · 1.46 KB

Intro

My approach to brain simulation

  1. Identify a list of computational modules in the brain
  2. For each module
    1. Understand what it computes: input -> output
    2. Get training data (synthetic or real-world)
    3. Train artificial neural networks to replicate its functionality
  3. Combine modules

List of implemented modules

pip install brain-modules
  1. Place Cells (incomplete, deprecated)
  2. Brain's GPS System (Head Direction Cells, Grid Cells, Place Cells)

List of implemented ANN modules

  1. Transformer (RMSNorm, RoPE, SwiGLU, GroupedQueryAttention)
  2. GRU (Gated Recurrent Unit)
  3. SwiGLUMLP (Swish-Gated Linear Unit MLP)