Logistic Regression: Deriving Gradients using Matrix Calculus
Derivation of $\frac{\partial L}{\partial M}$ for Cross Entropy Loss $L(y) = - \sum_{i=1}^C g\ log(y)$ with ground truth $g$, predicted $y$
Read more →
ML Foundations
13 Lessons
The Core of Backprop 3: Deriving $\frac{\partial L}{\partial M}$ using Matrix Calculus
Very elegant & short derivation of: $\frac{\partial L}{\partial M}$ for loss $L$ to update $M$ in Linear Layers
Read more →
Optimizing 2D Convolutions Implementation
Optimize Simple CNN Numpy Network to Denoise Image
Read more →
Optimizing 1D Convolutions Implementation
Simple Derivation to Optimize Kernel with Numpy Code to Denoise Signal
Read more →
Optimizing 1D Convolutions
Take the 1D Convolution as a Toeplitz Matrix & Optimize for Signal Denoising
Read more →
The Core of Backprop 2: Deriving $\frac{\partial L}{\partial M}$ using Matrix Calculus
Shorter **Matrix Calculus** Derivation of $\frac{\partial L}{\partial M}$ for loss $L$ to update $M$ in Linear Layers
Read more →
2D Convolutions Framed as Toeplitz Matrices
1D Convolutional Toeplitz Matrix $\rightarrow$ 2D Convolutional Toeplitz Matrix
Read more →
