IT and Tech Certifications
40 flashcards
~20 min
Artificial Intelligence and Machine Learning Basics
Learn the core concepts behind artificial intelligence and machine learning with this 40-card deck. It covers learning types, common algorithms, model training and evaluation, neural networks, and modern topics like transformers and large language models. It a...
About this deck
Learn the core concepts behind artificial intelligence and machine learning with this 40-card deck. It covers learning types, common algorithms, model training and evaluation, neural networks, and modern topics like transformers and large language models. It also touches on responsible AI. The deck suits beginners, tech students, and anyone preparing for entry-level AI or cloud certifications.
Ready to test yourself?
Flip through all 40 cards in interactive study mode.
All flashcards
Click any question to reveal the answer.
The field of computer science focused on building systems that perform tasks normally requiring human intelligence.
A subset of AI in which systems learn patterns from data instead of following explicitly programmed rules.
A branch of ML that uses neural networks with many layers to learn complex patterns.
Training a model on labeled data, where each input has a known correct output.
Finding hidden patterns or structure in unlabeled data.
A method where an agent learns by interacting with an environment and receiving rewards or penalties.
Training with a small amount of labeled data plus a large amount of unlabeled data.
An input variable used by a model to make predictions.
The correct output a model is trained to predict.
To train the model on one portion and evaluate it on unseen data to measure real-world performance.
Predicting which category an input belongs to, such as spam or not spam.
Predicting a continuous numeric value, such as a house price.
Grouping similar data points without labels. K-means is a common example.
A model that predicts a numeric value by fitting a straight line to the data.
A classification algorithm that outputs a probability, typically using the sigmoid function.
A model that makes predictions by following a series of if/else splits on features.
An ensemble of many decision trees whose predictions are combined, usually by voting or averaging.
An algorithm that finds the boundary (hyperplane) that separates classes with the maximum margin.
An algorithm that classifies a point based on the majority class of its k closest neighbors.
A probabilistic classifier based on Bayes' theorem that assumes features are independent.
When a model learns the training data too closely, including noise, and performs poorly on new data.
When a model is too simple to capture the patterns in the data and performs poorly even on training data.
The balance between a model that is too simple (high bias) and one that is too sensitive to training data (high variance).
Evaluating a model by training and testing it on several different splits of the data.
A technique, such as L1 or L2 penalties, that discourages model complexity to reduce overfitting.
A setting chosen before training, such as learning rate or tree depth, rather than learned from data.
A function that measures how far a model's predictions are from the true values.
An optimization algorithm that adjusts parameters step by step to minimize the loss.
A model made of layers of interconnected nodes (neurons) that learn patterns from data.
A function, such as ReLU, that adds non-linearity to a neuron's output so the network can learn complex patterns.
The algorithm that calculates the loss gradients for each weight so the network can update them.
One complete pass through the entire training dataset.
A neural network designed for grid-like data such as images, using filters to detect features.
A neural network designed for sequential data such as text or time series, which keeps a memory of earlier inputs.
A neural network architecture that uses attention to process sequences in parallel, introduced in 2017's "Attention Is All You Need."
The field of AI focused on enabling computers to understand and generate human language.
A very large transformer-based model trained on massive text data to understand and generate language.
The share of predicted positives that are actually positive: TP / (TP + FP).
The share of actual positives that the model correctly identifies: TP / (TP + FN).
Systematic unfair outcomes from an AI system, often caused by biased training data or flawed design.