AI Fundamentals: 20 Key Concepts

20 terms · by Sean · updated yesterday

Flashcards Flip through the deck and sort each card into “still learning” or “got it”. Learn Adaptive rounds — multiple choice first, then typing, weakest cards more often. Match Race the clock pairing terms with definitions. Six pairs a round. Test A graded mock exam: written, multiple choice and true/false, marked at the end. Crossword Terms become answers, definitions become clues. Solve it here or print it.

Terms in this set

Artificial Intelligence (AI)

The field of building computer systems that perform tasks normally requiring human intelligence, such as understanding language, recognizing patterns, reasoning, and making decisions.

Machine Learning (ML)

A subset of AI where systems learn patterns from data rather than being explicitly programmed with rules. Performance improves as the system sees more examples.

Deep Learning

A subset of machine learning that uses neural networks with many layers to learn increasingly abstract representations of data. It powers modern vision, speech, and language systems.

Neural Network

A model made of layers of interconnected nodes (neurons) that transform inputs into outputs. Each connection has a weight that is adjusted during training.

Supervised Learning

Training a model on labeled examples (inputs paired with correct outputs) so it can predict labels for new inputs. Used for classification and regression.

Unsupervised Learning

Training on unlabeled data to discover hidden structure, such as clusters or lower-dimensional representations. Examples: k-means clustering, PCA.

Reinforcement Learning (RL)

A learning approach where an agent takes actions in an environment and learns from rewards or penalties to maximize long-term reward.

Training vs. Inference

Training is the process of adjusting a model's parameters using data. Inference is using the trained model to make predictions on new inputs.

Parameters (Weights)

The numeric values inside a model that are learned during training. They determine how inputs are transformed into outputs. Large models have billions of them.

Loss Function

A measure of how far a model's predictions are from the correct answers. Training aims to minimize this value.

Gradient Descent

An optimization algorithm that repeatedly adjusts parameters in the direction that reduces the loss, using the gradient (slope) of the loss function.

Backpropagation

The algorithm that computes how much each weight in a neural network contributed to the error, by propagating gradients backward from output to input layers.

Overfitting

When a model memorizes training data, including its noise, and performs poorly on new data. Countered with more data, regularization, dropout, or early stopping.

Training, Validation, and Test Sets

Data is split three ways: the training set fits the model, the validation set tunes hyperparameters and guides model choice, and the test set gives a final unbiased performance estimate.

Transformer

A neural network architecture built on self-attention, which lets the model weigh the relevance of every part of the input to every other part. It is the foundation of modern language models.

Large Language Model (LLM)

A very large transformer-based model trained on huge amounts of text to predict the next token. This enables it to generate text, answer questions, write code, and more.

Token

The basic unit of text a language model processes: a word, part of a word, or a character. Models read and generate text one token at a time.

Embedding

A representation of data (words, images, etc.) as a vector of numbers, where similar items end up close together in the vector space.

Hallucination

When a generative AI model produces output that sounds plausible and confident but is false or unsupported by its sources.

AI Alignment

The field concerned with ensuring AI systems pursue goals and behave in ways consistent with human intentions and values, including being safe, honest, and beneficial.