7 terms
An activation function is the nonlinear rule a neural network neuron applies to its weighted input, letting stacked layers model complex patterns.
An attention mechanism lets a neural network weight the most relevant parts of its input for each output, and it is the core operation of Transformers.
Backpropagation is the algorithm that computes how much each weight in a neural network contributed to the error, so gradient descent can train it.
A convolutional neural network (CNN) is a neural network that learns spatial features with shared convolution filters, used mainly for images and signals.
A neural network is a machine learning model built from layers of weighted units that learns to map inputs to outputs by adjusting weights from data.
A neuron in a neural network is a computing unit that sums weighted inputs, adds a bias and applies an activation function to produce one output…
A transformer is a neural network architecture that uses self-attention to process sequences in parallel and underlies modern language and vision models.