AI 용어집
인공지능 완전 사전
Multilayer Perceptron
Fully connected neural network architecture composed of multiple hidden layers allowing to learn complex non-linear representations of input data.
Backpropagation
Optimization algorithm that calculates the gradients of the loss function with respect to each weight of the network by propagating the error from the output to the input.
Activation Function
Non-linear function applied to the output of each neuron to introduce non-linearity in the model and allow learning of complex representations.
Stochastic Gradient Descent
Optimization algorithm that updates the network weights using the gradient calculated on a single sample or mini-batch at each iteration.
Convolutional Neural Network
Architecture specialized in processing grid-like data such as images, using convolution layers to extract hierarchical features.
Recurrent Neural Network
Architecture designed to process sequential data by maintaining internal memory through recurrent connections between neurons.
Long Short-Term Memory
Advanced variant of RNNs using forget, input, and output gates to control information flow and solve the long-term dependency problem.
Vanishing Gradient Problem
Phenomenon where gradients become extremely small during backpropagation in deep networks, preventing effective learning of initial layers.