Add training algorithm terminology to ml.md

[Copilot]

Extends ml/ml.md with a "Training Algorithm Terms" section covering core optimization and training concepts.

Changes

• Added 9 training algorithm definitions: Loss Function, MSE, Optimizer, SGD, Learning Rate, Backpropagation, Gradients, Zero Gradients, and Step
• Follows existing formatting conventions with bold terms and concise explanations

The new section complements the existing general ML terminology with more specific training mechanics.

Original prompt

ADD THIS TO THE FILE

raining Algorithm Terms
Loss Function: Mathematical function that measures prediction error
MSE (Mean Squared Error): Loss function that squares the difference between predictions and actual values
Optimizer: Algorithm that updates model parameters to reduce loss
SGD (Stochastic Gradient Descent): Simple optimizer that updates parameters using gradients
Learning Rate: How much to change parameters each step (0.01 = small, careful steps)
Backpropagation: Algorithm to compute gradients of loss with respect to all parameters
Gradients: Partial derivatives showing how loss changes with parameter changes
Zero Gradients: Clearing old gradient values before computing new ones
Step: The optimizer's update operation (moving parameters in the right direction)

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[senthil4321]

All ok