Skip to Main Content Carnegie Mellon University Libraries

Artificial Intelligence Research: Home

Introduction

This guide introduces the key artificial intelligence (AI) glossary of concepts. Under each tab, you can find additional resources from CMU Libraries about AI-related research data and articles.

AI Glossary

Artificial Intelligence: A field of study that researches and develops intelligent machines capable of simulating human intelligence.

Algorithm: Instructions defining a set of tasks that a computer must follow sequentially to learn from data.

Chatbot: A computer program that can simulate a conversation with a human end user, using techniques such as NLP to parse inputs (or questions) and generative AI to provide outputs (or answers).

Computer Vision: The extraction of information from visual data. Includes such techniques as contextual image classification, facial recognition, object detection, and image segmentation. 

Deep Learning: A machine learning method by which computers learn in a way that mimics the human brain, through analyzing large amounts of information and classifying that information into discrete categories. Deep learning algorithms are based on neural networks.

Generative AI: Deep learning models that process raw data (image or text) and then generate statistically likely outputs based on that input raw data. Generative AI applications are increasingly “foundation models”—trained on large swaths of unlabeled data so that they can be used in a wide range of applications and domains.

Large Language Model: A program trained on swaths of textual data that uses Natural Language Processing to recognize statistical patterns and semantic relationships between words. The recognition and repetition of these patterns allow it to create human-sounding outputs and responses, such as those generated by Bard, CoPilot, and ChatGPT.

Machine Learning: The process of machines using algorithms to adapt and improve their perception, cognition, and action without being programmed to do so, based on inputs such as data, knowledge, experience, and interaction.

Natural Language Processing: A machine learning technique that allows computers to process and interpret human language through text or voice data.

Neural Network: A subset of machine learning and the basis for deep learning algorithms, neural networks work to mimic neurons connecting and signaling to each other in the human brain.

Supervised Learning: A machine learning method that uses labeled data to train a model that predicts the output.

Training Data: Data used to train machine learning models.

Testing Data: Data used to assess the accuracy of machine learning models.

Unsupervised Learning: A machine learning method where the model learns from unlabeled training data to explore patterns without guidance.

Engineering Librarian