Understanding AI: A Complete A-Z Terminology

Here is a dictionary of A-Z terms which are related ( or you might have heard of ) in context of AI

A — Artificial Intelligence
B — Bias
C — ChatGPT
D — Dall-E/Dataset/Dense Vectors
E — Embeddings
F — Fine Tuning
G — Generative AI
H — Hugging Face
I — Inference
J — Jyupter Notebooks
K — Knowledge Graph
L — Large Language Models
M — Machine Learning
N — Neural Networks
O — Overfitting
P — Prompts
Q — Quantization
R — Retrieval Augmented Generation (RAG)
S — Supervised Learning
T — Tokens
U — Unsupervised Learning
V — Vector Database
W — Weights
X — Xavier Initialisation
Y — Yield
Z — Zero shot learning

A — Artificial Intelligence
“Artificial Intelligence is when machines can think and learn like humans.”
Example: AI can help recommend movies you might like based on what you’ve watched before.

B — Bias
“Bias in AI means that the machine can make unfair decisions based on its training data.”
Example: If an AI learns from biased data, it might favor one group over another in job applications.

C — ChatGPT
“ChatGPT is a computer program that can chat with you like a real person.”
Example: You can ask ChatGPT questions, and it will give you answers in a friendly way.

D — Dall-E
“Dall-E creates pictures from words, letting you see your ideas come to life.”
Example: You can say “a dog in a superhero costume,” and Dall-E will draw it for you.

E — Embeddings
“Embeddings turn words into numbers so machines can understand their meanings.”
Example: The words “king” and “queen” will have similar numbers because they are related.

F — Fine Tuning
“Fine tuning is adjusting a model to make it better at a specific task.”
Example: A general AI can be fine-tuned to understand medical terms for healthcare use.

G — Generative AI
“Generative AI can create new things, like text, images, or music.”
Example: It can write a poem or compose a song based on a few words you give it.

H — Hugging Face
“Hugging Face is a place where people share AI tools and models.”
Example: You can find many pre-built AI models there to use in your projects.

I — Inference
“Inference is when an AI makes predictions or decisions based on what it has learned.”
Example: After training, an AI can predict if an email is spam or not.

J — Jupyter Notebooks
“Jupyter Notebooks are tools where you can write and run code while seeing results right away.”
Example: Data scientists use Jupyter to test their code and visualize data easily.

K — Knowledge Graph
“A knowledge graph connects different pieces of information, helping machines understand relationships.”
Example: Google uses a knowledge graph to show related facts when you search for something.

L — Large Language Models
“Large language models are powerful AI that understand and generate human-like text.”
Example: GPT-3 is a large language model that can write stories and answer questions.

M — Machine Learning
“Machine learning is a way for computers to learn from data without being told exactly what to do.”
Example: A machine learning model can learn to recognize cats in photos by analyzing many images.

N — Neural Networks
“Neural networks are computer systems that work like the human brain to process information.”
Example: They are used in image recognition to identify objects in pictures.

O — Overfitting
“Overfitting happens when a model learns too much from its training data and doesn’t work well on new data.”
Example: If a model memorizes answers instead of learning patterns, it can fail on different questions.

P — Prompts
“Prompts are the instructions you give to AI to get a response.”
Example: Asking, “Tell me a joke” is a prompt that makes the AI respond with humor.

Q — Quantization
“Quantization makes AI models smaller and faster so they can run on devices like phones.”
Example: A quantized model can perform tasks quickly without needing a powerful computer.

R — Retrieval Augmented Generation (RAG)
“RAG combines searching for information and generating answers to make responses better.”
Example: It can look up facts while answering your question to provide accurate information.

S — Supervised Learning
“Supervised learning is when a model learns from labeled examples to make predictions.”
Example: Teaching an AI to recognize fruits by showing it pictures of apples and oranges with labels.

T — Tokens
“Tokens are pieces of text that AI uses to understand and generate language.”
Example: The sentence “I love ice cream” can be broken down into tokens like “I,” “love,” “ice,” and “cream.”

U — Unsupervised Learning
“Unsupervised learning is when a model finds patterns in data without labels.”
Example: It can group similar customers based on their shopping habits without knowing their names.

V — Vector Database
”A vector database stores information in a way that makes

W — Weights

“Weights are numbers that determine how much influence each input has on the model’s output.”
Example: In a neural network, weights are adjusted during training to improve predictions, like deciding how important different features of an image are for identifying it.

X — Xavier Initialization

“Xavier Initialization is a method used to set the starting values of weights in a neural network.”
Example: By using Xavier Initialization, we help the model learn better and faster by keeping the weights balanced at the beginning of training.

Y — Yield

“Yield refers to the output or results produced by an AI model after processing data.”
Example: If you input data into a model, the yield would be the predictions or classifications it provides based on that data.

Z — Zero-Shot Learning

“Zero-shot learning is when a model can make predictions about things it has never seen before.”
Example: If an AI trained to recognize animals is asked to identify a zebra without ever having seen one, it can do so by using its understanding of similar animals, like horses and stripes.