What Is an Optimizer in Machine Learning?
An optimizer is the procedure that changes model parameters to reduce a training objective such as loss.
Don't Just Read it – Explore It
An optimizer is the procedure that changes model parameters to reduce a training objective such as loss.
A perceptron is a simple linear classifier that combines input features, applies weights, and predicts one of two classes.
A multilayer perceptron is a feedforward neural network with an input layer, one or more hidden layers, and an output layer.
Algorithmic fairness examines how an automated system’s benefits, errors, and burdens are distributed across people and contexts.
An activation function transforms a neuron’s weighted input, allowing a neural network to represent nonlinear relationships.
Prompt injection is an attack that places untrusted instructions into model input to override or redirect higher-trust instructions.
Instruction tuning fine-tunes a language model on tasks expressed as natural-language instructions and desired responses.
Reinforcement learning from human feedback uses human comparisons to train a model toward preferred behavior.
Monte Carlo tree search grows a decision tree by balancing promising choices with exploration and using simulated outcomes.
Fine-tuning continues training a pretrained model on a smaller dataset to adapt its behavior to a task or domain.
Q-learning estimates the long-term value of taking an action in a state and then following the best learned choices.
A graph neural network learns representations by combining information from connected nodes and edges.