Deep Learning
Architectures
CNNs, attention, and transformers — what each structure buys and why it won
Attention mechanism
coreScaled dot-product attention as a differentiable soft lookup — queries score keys, softmax turns scores into weights, the output is a weighted mix of values
12 minarchitecturesattentiontransformersConvolutional neural networks
coreLocality and weight sharing turn images into a parameter-efficient feature hierarchy — plus the output-size and parameter-count formulas interviewers make you compute
11 minarchitecturescnncomputer-visionRNNs & LSTMs
coreVanilla recurrence and backpropagation through time, why gradients vanish or explode across long sequences, and how gated cells (LSTM, GRU) restore long-range learning
12 minarchitecturesrnnlstmsequence-modelsTransformer architecture
advancedMulti-head attention plus position-wise FFNs, residuals, and positional encodings — and the parallelism argument that retired recurrent networks
12 minarchitecturestransformersattention