https://arunprakaash.github.io/lab/neural-nets/ Recent content in Neural Nets on arun._space Hugo en-us Sat, 02 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-05-backpropagation/ Sat, 02 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-05-backpropagation/ Loss functions, gradient descent, and backpropagation — how a neural network looks at its mistakes and figures out exactly who to blame. https://arunprakaash.github.io/lab/neural-nets/lesson-06-training-loop/ Sat, 02 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-06-training-loop/ Forward pass, backprop, gradient descent — assembled into a training loop. Watch a network learn to separate XOR, circles, and spirals in real time. https://arunprakaash.github.io/lab/neural-nets/lesson-07-depth-problem/ Sat, 02 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-07-depth-problem/ Why deep networks went dark in the 90s — vanishing gradients, exploding gradients, and the tricks that finally made depth work: weight init, batch norm, and residual connections. https://arunprakaash.github.io/lab/neural-nets/lesson-08-from-scratch/ Sat, 02 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-08-from-scratch/ Stop using the concepts, start writing the code. A full neural network in pure Python and NumPy — same thing PyTorch does internally, just slower. https://arunprakaash.github.io/lab/neural-nets/lesson-09-overfitting/ Sat, 02 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-09-overfitting/ When your network aces the training data and fails at everything else — overfitting, regularisation, dropout, and how to actually tell if your model is learning. https://arunprakaash.github.io/lab/neural-nets/lesson-10-cnns/ Sat, 02 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-10-cnns/ How convolutional neural networks see images — kernels, feature maps, pooling, and why a sliding 3×3 window beats a million fully-connected weights. https://arunprakaash.github.io/lab/neural-nets/lesson-01-math-foundations/ Fri, 01 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-01-math-foundations/ Before we touch a single neuron, we need to speak its language. Vectors, matrices, derivatives, and the chain rule — the math that makes neural networks tick. https://arunprakaash.github.io/lab/neural-nets/lesson-02-perceptron/ Fri, 01 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-02-perceptron/ One neuron. A handful of weights. A rule so simple it fits in one line. And yet — it learns. This is the perceptron, and it’s where everything begins. https://arunprakaash.github.io/lab/neural-nets/lesson-03-activation-functions/ Fri, 01 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-03-activation-functions/ Sigmoid, ReLU, tanh, softmax — the nonlinear magic that makes depth actually mean something. https://arunprakaash.github.io/lab/neural-nets/lesson-04-forward-pass/ Fri, 01 May 2026 00:00:00 +0000 https://arunprakaash.github.io/lab/neural-nets/lesson-04-forward-pass/ The forward pass: how a number enters one end of a neural network and a prediction falls out the other, layer by layer.