Artificial Intelligence Supervised Learning

Flashcards summarizing key vocabulary and concepts from the lecture on Artificial Intelligence and Supervised Learning.

Attributes

Inputs or features in supervised learning, include targets and outputs.

Model type

The architecture or method used to produce predictions from inputs.

Weights

Real-valued coefficients in a model that determine the influence of input features.

Objective function

A function that the learning algorithm seeks to minimize, usually related to error.

Input attributes

Variables represented by a vector, typically used to describe the input data.

Supervised learning task

Learning a function that approximates an unknown function from a training set of input-output pairs.

Ground truth

The actual labels or outputs that serve as a reference for the model's predictions.

Test set

A separate set of (x, y) pairs used to evaluate the model's performance.

Overfitting

When a model performs well on training data but poorly on unseen test data.

Underfitting

When a model fails to capture the underlying trend of the data, resulting in poor performance on both training and test data.

Simple Linear Regression

A method to model the relationship between a single input variable (x) and an output variable (y) through a linear equation.

Least squares

A method for minimizing the sum of the squares of the differences between observed and predicted values.

Gradient descent

An optimization algorithm used to minimize the loss function by iteratively moving in the direction of the steepest descent.

Activation function

A function that determines the output of a neural network node based on its input.

Computation graph

A representation of the operations and data flows in a neural network, highlighting how inputs progress through layers.

Logistic function

A sigmoid function that provides the probability estimate of a binary classification.

One-vs-rest classifier

A classification strategy where separate classifiers are trained to distinguish one class from all others.

One-vs-one classifier

A strategy where classifiers are trained for each pair of classes, suitable for multi-class classification.

Learning curve

A plot that illustrates how model performance improves with increasing amounts of training data.