What, Where & How: A Glossary Of Terms Related To Artificial Intelligence, Machine Learning and Deep Learning

A glossary of terms is always a work in progress, given how technology, methods, and techniques evolve.

This is more true for the fields of Artificial Intelligence, Machine Learning, and Deep Learning than others, which is why our developers have put together a list of more commonly used terms to help you tell your Algorithms from your Active Learning, and Selection Bias from Sentiment Analysis.

Artificial Intelligence
A field of computer engineering that focuses on creating systems capable of gathering data and making decisions and/or solving problems.

Active Learning
This is a training approach in which an algorithm chooses some of the data it learns from. An active learning algorithm selectively seeks the particular range of examples it needs for learning, rather than blindly seeking a diverse range of labeled examples. Active learning is particularly valuable when labeled examples are scarce.

Algorithm
Math formulas and/or programming commands that inform non-intelligent computers how to solve problems using AI.

Artificial Neural Network (ANN)
A model used in AI based on the human brain that consists of neural layers.

Bayesian Neural Network
This is a probabilistic neural network that relies on Bayes’ Theorem to calculate uncertainties in weights and predictions. It predicts a distribution of values and can be useful when it is important to quantify uncertainty.

Bucketing
This refers to converting a feature into multiple binary features called buckets or bins, based on value range.

Classification Model
This is a type of Machine Learning model for distinguishing among two or more discrete classes.

Clustering
This refers to the grouping of related examples, particularly during unsupervised learning. A human can optionally supply meaning to each cluster once all the examples are grouped.

Deep Learning
This is a Machine Learning method that consists of a many-layered Artificial Neural Network and uses many layers of nonlinear processing to extract features from the data before transforming it into different levels of abstraction.

Data Analysis
This refers to how an understanding of data is obtained by considering samples, measurement, and visualization. It is crucial in understanding experiments and debugging problems with the system.

Data Augmentation
This is when the range and number of training examples are artificially boosted by transforming existing examples to create additional examples. It can produce many variants of the original example, possibly yielding enough labeled data to enable training.

Dataset
A collection of examples.

Decision Tree
A model represented as a sequence of branching statements.

Dynamic Model
A model trained online in a continuously updating fashion.

Heuristics
Common-sense rules based on experience.

Input Layer
The first layer that receives the input data in a neural network.

Layer
A set of neurons in a neural network that processes a set of input features, or the output of those neurons.

Metric
A number that you care about, that may or may not be directly optimized in a Machine Learning system.

Machine Learning
A subset of AI that uses algorithms to learn from identified patterns in data, then adjusts actions accordingly, without explicit programming.

Neural Network
A network designed to be similar to the human nervous system and brain that gives AI the ability to solve complex problems.

Natural Language Processing
AI trained to interpret human communication.

Objective
A metric that an algorithm is trying to optimize.

Reinforcement Learning
A method of teaching a machine that involves running scenarios and reporting results, then using the feedback to achieve better results.

Supervised Learning
A learning method that provides the machine with the correct answer ahead of time.

Selection Bias
This refers to errors in conclusions drawn from sampled data due to a process that generates systematic differences between samples observed in the data and those not observed.

Sentiment Analysis
The use of statistical or Machine Learning algorithms to determine a group’s overall attitude, be it positive or negative, toward a service, product or organization.

Unsupervised Learning
Feeding the machine data and allowing it to find whatever patterns it is able to.

Turing Test
Developed by Alan Turing in 1950, this tests a machine’s ability to behave in an intelligent manner that is indistinguishable from human behavior.

For more information on any or all of these terms, or how Artificial Intelligence, Machine Learning, and Deep Learning can help your business, talk to us about our award-winning CRM solutions and Intelligent Customer Management platform today.

Author: Adam Edmonds

Vice President, Products

With almost 2 decades of experience developing customer management solutions in financial services and insurance, Adam Edmonds is responsible for establishing overall product vision and designing easy to use solutions that solve real market problems.

Adam is excited to share the lessons he has learned and his insights on where the industry is heading with readers of his blog. He encourages readers to join the discussion or reach out to him with their own insights, best practices, and solutions to industry challenges.