Common ML & AI terms and definitions

Patrick Mazzotta
Co-Founder 16 May, 2024

Welcome to our quick guide on machine learning (ML) and artificial intelligence (AI) terminology! We’ve had a few conversations with people about all the buzz terms and context around them and thought it would be a good idea to write up a short reference1.

So whether you’re just starting out or looking to brush up on your a few specific terms, we hope this post is helpful! Feel free to bookmark this page and come back to it whenever you need!

Artificial Intelligence (AI): A broad area of computer science creating systems that perform tasks requiring human intelligence, such as learning and problem-solving.
Machine Learning (ML): A subset of AI using algorithms that enable machines to learn from data and make predictions or decisions.
Model: The output of an ML algorithm after training, representing what the algorithm has learned from data.
Natural Language Processing (NLP): AI technology enabling computers to understand and respond to human language effectively.
Large Language Model (LLM): Advanced AI models capable of understanding and generating human-like text, often used for tasks like translation, summarization, and question-answering.
Generative Pre-trained Transformer (GPT): A type of language processing AI model known for its ability to generate coherent and contextually relevant text based on input.
Drift: In ML and AI, the phenomenon where a model’s performance degrades over time due to changes in the underlying data or environment.
Hallucination: In generative AI, especially with GPT, the tendency for the AI to produce incorrect or nonsensical information that seems plausible.
Bias: In ML, bias refers to systematic errors in the model due to incorrect assumptions in the learning algorithm or the data.
Supervised Learning: A type of ML where the model is trained on labeled data, learning to predict outputs from given inputs.
Unsupervised Learning: A type of ML where the model is trained on unlabeled data, discovering hidden patterns without explicit instructions.
Neural Network: A series of algorithms modeled after the human brain that recognizes patterns and interprets data.
Deep Learning: A subset of ML involving neural networks with many layers, enabling the processing of large amounts of data to identify complex patterns.
Reinforcement Learning: A type of ML where an agent learns to make decisions by receiving rewards or penalties for actions taken.
Feature Engineering: The process of selecting, modifying, and creating variables (features) that help ML models perform better.
Overfitting: When a model learns the training data too well, including noise and outliers, resulting in poor performance on new, unseen data.
Underfitting: When a model is too simple to capture the underlying patterns in the data, leading to poor performance on both training and new data.
Generative Adversarial Networks (GANs): A class of ML frameworks where two neural networks, a generator and a discriminator, compete against each other. The generator creates data while the discriminator evaluates it, leading to the generation of highly realistic data.

We hope this helps you get comfortable with the core concepts in AI & ML. If you’re looking for more specific support or advice on machine learning & artificial intelligence, feel free to contact us. We love being helpful and are willing and able to help you navigate the complexities of artificial intelligence & machine learning.


  1. Disclaimer: we’re not a dictionary and these definitions are based on our hands-on experience with ML & AI informed by the perspective our team has working in this space. We think everything here is correct, but we accept and expect some people might use different definitions or challenge some of our positioning.
Written by Patrick Mazzotta

Patrick is the co-founder and head of data sciences at Ones & Heroes. He also enjoys designing quant-based investment strategies and can be frequently found immersed in VR games.