Synthesis — Cross-Topic Conceptual Maps

These documents connect ideas across multiple topics. They answer questions like: - When should I use model A vs model B? - How do loss functions relate across models? - What's the geometric view of ML?

Documents

Document	Connects
optimization_methods_compared.md	GD, Newton, closed-form across models
loss_functions_map.md	MSE, CE, hinge, log-loss
bias_variance_tradeoff.md	Fundamental tradeoff across all models
geometry_of_ml.md	Geometric view: regression, PCA, SVM
probabilistic_view_of_ml.md	MLE, MAP, Bayesian connections
model_selection_guide.md	Decision framework
supervised_vs_unsupervised.md	Taxonomy and connections
regularization_across_models.md	L1/L2/dropout/early stopping
deep_learning_building_blocks.md	Inductive biases of MLP, CNN, RNN, attention, AE, GNN
sequence_models_and_attention.md	The arc RNN → LSTM → Transformer → LLM engineering
generative_and_self_supervised.md	VAE, GAN, diffusion, MAE, InfoNCE, RLHF/DPO

Format

All synthesis documents are .md (pure text). They reference topics/ notebooks for computational details.