Glossary of AI terms

Artificial intelligence is the process of programming a computer to make decisions for itself. This can be done in several ways, but the most common is to use algorithms, or sets of rules, to sort and process data.

Artificial intelligence can be used for several tasks, from simple pattern recognition to more complex tasks such as decision-making and natural language processing. Because artificial intelligence can be used to automate decision-making, it has the potential to greatly improve efficiency in many different areas of life.

For example, it can be used to help doctors diagnose diseases, plan military campaigns, or even drive cars. As artificial intelligence continues to develop, its potential applications are only limited by our imagination.

As artificial intelligence technology advances, it will become increasingly important for people to understand the basics of AI. This is because AI will play an increasingly important role in our lives, from the way we work to the way we socialize.

AI is already being used in many areas, such as finance, healthcare, and manufacturing. In the future, it will likely be used in more areas, such as transportation and agriculture. As such, everyone needs to have a basic understanding of what AI is and how it works. This will help them make better decisions about how they want to use this technology in their own lives.

When it comes to artificial intelligence, there are a lot of terms and concepts that can be confusing for those who are new to the topic. Let’s dive into some of the most common AI terms.

Featured AI Terms

Glossary of AI terms

Accuracy: Accuracy is a measure used in binary classification to determine the correctness of an answer or output. It is calculated by adding the number of true positives and true negatives and then dividing it by the sum of true positives, true negatives, false positives, and false negatives.
Algorithm: A set of rules or steps that are followed in order to solve a problem.
Artificial intelligence: A field of computer science that deals with the development of intelligent machines that can perform tasks that would normally require human intelligence, such as understanding natural language and recognizing objects.
Autonomous: Able to operate independently. In the context of AI, autonomous systems are able to make their own decisions without human input.

Backward chaining: A reasoning method in which inferences are made from a given goal back to the data or information that is needed in order to achieve the goal.
Bias: A prejudice or preference for one thing over another. Bias can be positive (favoring one thing) or negative (against one thing). In machine learning, bias refers to the error introduced by oversimplifying a model.
Big data: A term used to describe data sets that are too large and complex to be processed by traditional methods.
Bounding box: A type of annotation used in computer vision that defines a rectangle around an object in an image.

Chatbot: A computer program that simulates human conversation, usually via text or voice messages. Chatbots are commonly used as virtual customer service representatives.
Cognitive computing: A subfield of artificial intelligence that deals with the development of systems that can reason, learn, and solve problems like humans.
Computational learning theory: A branch of machine learning that deals with the mathematical properties of learning algorithms.
Corpus: A collection of texts, usually in digital form, that is used for linguistic research.

Data mining: The process of extracting patterns from large data sets.
Data science: A field that deals with the extraction of knowledge from data.
Dataset: A collection of data that is used for training a machine learning model.
Deep learning: A subfield of machine learning that deals with the development of neural networks that can learn from data without being explicitly programmed.

Entity annotation: The process of identifying and labeling entities in text data. Entity annotation is often used for named entity recognition.
Entity extraction: The process of identifying and extracting entities from text data. Entity extraction is often used for named entity recognition.

Forward chaining: A reasoning method in which inferences are made from given data or information forward to the desired goal.

General AI: A term used to describe artificial intelligence systems that exhibit general intelligence, that is, the ability to reason and solve problems like humans.

Hyperparameter: A parameter that is not directly learned by a machine learning model but is instead set prior to training. The values of hyperparameters can have a significant impact on the performance of a model.

Intent: In the context of chatbots and natural language understanding, intent refers to the goal or purpose of an utterance. For example, the intent of the utterance “I’m looking for a hotel” may be to find a hotel.

Label: A category or class that a data point can be assigned to. In supervised learning, labels are used to train models.
Linguistic annotation: The process of adding linguistic information to text data. Linguistic annotation is often used for named entity recognition and sentiment analysis.

Machine intelligence: A term used to describe artificial intelligence systems that exhibit human-like intelligence, that is, the ability to reason and solve problems like humans.
Machine learning: A subfield of artificial intelligence that deals with the development of algorithms that can learn from data without being explicitly programmed.
Machine translation: The process of translating text from one language to another using a machine.
Model: A mathematical representation of a real-world system. Models can be used to make predictions based on data.

Neural network: A computer system that is modeled after the brain and nervous system. Neural networks are commonly used in deep learning.
Natural language generation (NLG): The process of generating text from meaning representations. NLG is often used to generate summaries or descriptions from data.
Natural language processing (NLP): The process of understanding and manipulating natural language data. NLP is often used for tasks such as named entity recognition and sentiment analysis.
Natural language understanding (NLU): The process of understanding natural language data. NLU is often used for tasks such as named entity recognition and sentiment analysis.
Named entity recognition (NER): The task of identifying and labeling entities in text data. NER is often used to extract information from unstructured text.

Overfitting: The process of developing a model that fits the training data too closely. Overfitting can lead to poor performance on test data.

Parameter: A value that is learned by a machine learning model during training. Parameters are used to make predictions.
Pattern recognition: The process of recognizing patterns in data. Pattern recognition is often used for classification tasks.
Predictive analytics: The process of using data to make predictions about future events. Predictive analytics is often used for marketing and fraud detection.
Python: A programming language that is widely used in artificial intelligence and machine learning.

Reinforcement learning: A subfield of machine learning that deals with the development of algorithms that can learn from data by trial and error.

Semantic annotation: The process of adding semantic information to text data. Semantic annotation is often used for named entity recognition and sentiment analysis.
Sentiment analysis: The task of identifying the sentiment of a text document. Sentiment analysis is often used to gauge customer opinion on a product or service.
Strong AI: A term used to describe artificial intelligence systems that exhibit human-like intelligence, that is, the ability to reason and solve problems like humans.
Supervised learning: A subfield of machine learning that deals with the development of algorithms that can learn from labeled data.

Test data: Data that is used to evaluate a machine learning model. Test data is usually held out from the training data.
Training data: Data that is used to train a machine learning model. Training data is usually split into training and test sets.
Transfer learning: The process of applying knowledge from one domain to another. Transfer learning is often used to improve the performance of machine learning models.
Turing test: A test for determining whether or not a machine can exhibit human-like intelligence. The Turing test was proposed by Alan Turing in 1950.

Unsupervised learning: A subfield of machine learning that deals with the development of algorithms that can learn from unlabeled data.

Validation data: Data that is used to evaluate a machine learning model. Validation data is usually held out from the training data.
Variance: Models with high variance are often difficult to use because they’re prone to both overfitting and low predictive accuracy.
Variation: These work in tandem with intents for natural language processing. Variation is what a person might say to achieve certain goals or purposes, and words can have multiple meanings when used this way!

Weak AI: A term used to describe artificial intelligence systems that are designed for specific tasks and cannot exhibit human-like intelligence.