What is a vector?
Very overloaded term whose meaning varies across different mathematical and scientific fields.
vector explained in plain English
Very overloaded term whose meaning varies across different mathematical and scientific fields. Within machine learning, a vector has two properties: - Data type: Vectors in machine learning usually hold floating-point numbers. - Number of elements: This is the vector's length or its dimension. For example, consider a feature vector that holds eight floating-point numbers. This feature vector has a length or dimension of eight. Note that machine learning vectors often have a huge number of dimensions. You can represent many different kinds of information as a vector. For example: - Any position on the surface of Earth can be represented as a 2-dimensional vector, where one dimension is the latitude and the other is the longitude. - The current prices of each of 500 stocks can be represented as a 500-dimensional vector. - A probability distribution over a finite number of classes can be represented as a vector. For example, a multiclass classification system that predicts one of three output colors (red, green, or yellow) could output the vector`(0.3, 0.2, 0.5)` to mean`P[red]=0.3, P[green]=0.2, P[yellow]=0.5`. Vectors can be concatenated; therefore, a variety of different media can be represented as a single vector. Some models operate directly on the concatenation of many one-hot encodings. Specialized processors such as TPUs are optimized to perform mathematical operations on vectors. A vector is a tensor of rank 1.
Example
Practitioners refer to vector when building, training, or evaluating machine learning systems. It appears in research papers, product documentation, and technical discussions about AI capabilities and limitations.
People also read
- AUC
A number between 0.
- Backpropagation
The process that tells a neural network which internal settings caused an error and how to adjust them, working backwards through layers.
- Bayesian neural network
A probabilistic neural network that accounts for uncertainty in weights and outputs.
- Bayesian optimization
A probabilistic regression model technique for optimizing computationally expensive objective functions by instead optimizing a surrogate that quantifies the uncertainty using a Bayesian learning technique.
- classification threshold
In a binary classification, a number between 0 and 1 that converts the raw output of a logistic regression model into a prediction of either the positive class or the negative class.
- configuration
The process of assigning the initial property values used to train a model, including: hyperparameters such as: - learning rate - iterations - optimizer - loss function In machine learning projects, c
- confusion matrix
An NxN table that summarizes the number of correct and incorrect predictions that a classification model made.
- cross-entropy
A generalization of Log Loss to multi-class classification problems.
- discriminative model
A model that predicts labels from a set of one or more features.
- embedding layer
A special hidden layer that trains on a high-dimensional categorical feature to gradually learn a lower dimension embedding vector.