#machine-learning

Coisas que escrevi sobre #machine-learning

Attention Mechanism

The attention mechanism allows neural networks to selectively focus on relevant parts of the input when producing each output — the core idea behind Transformers and modern LLMs.

Gradient Descent

Gradient descent is the backbone of training neural networks. This note covers the core algorithm, its variants (SGD, Mini-batch, Adam), and practical tips for using them effectively.

Mathematical Optimization

Mathematical optimization is a way of formally describing decision problems and solving them using dedicated algorithms. Models consist of decision variables, an objective function, and constraints that together define the best possible solution.

Regularization in ML

Regularization techniques prevent overfitting by constraining model complexity. This note covers L1, L2, Dropout, Batch Normalization, and early stopping with practical code examples.

CNNs: Uncovering Visual Patterns

Entenda como as Redes Neurais Convolucionais (CNNs) revolucionaram a visão computacional ao aprender hierarquias de características visuais.

Hierarchical Clustering

Explore o Clustering Hierárquico, um método poderoso para agrupar dados e entender suas relações hierárquicas, com exemplos e aplicações práticas.

BERT

Bidirectional Encoder Representations from Transformers.