andy
23991f92c9
Affected files: .obsidian/app.json .obsidian/workspace-mobile.json .obsidian/workspace.json Events/🪣🪣🪣.md STEM/AI/Neural Networks/CNN/FCN/README.md STEM/AI/Neural Networks/CNN/GAN/README.md STEM/AI/Neural Networks/CNN/README.md STEM/AI/Neural Networks/MLP/README.md STEM/AI/Neural Networks/README.md STEM/AI/Neural Networks/RNN/README.md STEM/AI/Neural Networks/SLP/README.md STEM/AI/Neural Networks/Transformers/README.md Untitled.canvas |
||
|---|---|---|
| .. | ||
| Back-Propagation.md | ||
| Decision Boundary.md | ||
| MLP.md | ||
| README.md | ||
- Architectures
- Single hidden layer can learn any function
- Universal approximation theorem
- Each hidden layer can operate as a different feature extraction layer
- Lots of Weight Init to learn
- Back-Propagation is supervised
!
Universal Approximation Theory
A finite Architectures MLP with 1 hidden layer can in theory approximate any mathematical function
- In practice not trainable with Back-Propagation
!
!
Weight Matrix
- Use matrix multiplication for layer output
- TLU is hard limiter
!
o_1too_4must all be one to overcome -3.5 bias and force output to 1 !
- Can generate a non-linear Decision Boundary