🤖

AI & Machine Learning Concepts

97 cards · 🇺🇸 English

Core AI/ML vocabulary: supervised vs unsupervised learning, neural networks, deep learning.

technologyaimachine-learning

Study all 97 cards for free

Flashcards, quizzes, pixel-art games and a daily review that remembers for you.

Study this deck free

Try the first 10 cards

1 / 10
Question

Artificial Intelligence (AI)

Knew itLearning

Tap the card for the answer · swipe right if you knew it, left if not

Already have an account? Sign in

Cards

  1. 1Artificial Intelligence (AI)
    The capability of computational systems to perform tasks typically associated with human intelligence: learning, reasoning, problem-solving, perception, decision-making
  2. 2Machine Learning
    Study of programs that can improve their performance on a given task automatically, without being explicitly programmed
  3. 3Deep Learning
    Type of machine learning that runs inputs through biologically inspired artificial neural networks with multiple layers
  4. 4Supervised Learning
    Machine learning where training data is labeled with expected answers; includes classification and regression
  5. 5Unsupervised Learning
    Machine learning that finds patterns and makes predictions in unlabeled data without guidance
  6. 6Reinforcement Learning
    Agent learns by being rewarded for good responses and punished for bad ones; learns optimal actions through trial and error
  7. 7Neural Network
    Computing system inspired by biological neural networks; interconnected nodes (neurons) process information in layers
  8. 8Artificial Neuron
    Basic unit of neural network; receives inputs, applies weights and activation function, produces output
  9. 9Activation Function
    Function in neural network that determines neuron output; introduces non-linearity (e.g., ReLU, sigmoid, tanh)
  10. 10Backpropagation
    Algorithm for training neural networks; calculates gradient of loss function and adjusts weights to minimize error
  11. 11Convolutional Neural Network (CNN)
    Deep learning architecture specialized for processing grid-like data (images); uses convolutional layers to detect features
  12. 12Recurrent Neural Network (RNN)
    Neural network with loops allowing information persistence; processes sequential data (text, time series)
  13. 13LSTM
    Long Short-Term Memory; type of RNN that can learn long-term dependencies; solves vanishing gradient problem
  14. 14Transformer
    Deep learning architecture using attention mechanisms; revolutionized NLP; basis for GPT, BERT models (introduced 2017)
  15. 15Attention Mechanism
    Allows model to focus on relevant parts of input when processing; key component of transformers
  16. 16Transfer Learning
    Using knowledge gained from solving one problem and applying it to a different but related problem
  17. 17Fine-tuning
    Process of taking pre-trained model and adapting it to specific task by training on smaller, task-specific dataset
  18. 18Overfitting
    Model performs well on training data but poorly on new data; memorizes instead of learning general patterns
  19. 19Underfitting
    Model too simple to capture underlying patterns; performs poorly on both training and test data
  20. 20Regularization
    Techniques to prevent overfitting by adding constraints or penalties (e.g., L1, L2 regularization, dropout)
  21. 21Dropout
    Regularization technique randomly dropping neurons during training to prevent overfitting
  22. 22Loss Function
    Measures how well model predictions match actual values; optimization aims to minimize loss
  23. 23Gradient Descent
    Optimization algorithm that iteratively adjusts parameters in direction that reduces loss function
  24. 24Stochastic Gradient Descent (SGD)
    Gradient descent variant using random subsets (mini-batches) of data instead of entire dataset; faster, more scalable
  25. 25Learning Rate
    Hyperparameter controlling step size in gradient descent; affects convergence speed and quality
  26. 26Hyperparameter
    Parameter set before training (not learned); e.g., learning rate, batch size, number of layers
  27. 27Cross-Validation
    Technique splitting data into multiple folds to evaluate model performance; reduces overfitting risk
  28. 28K-Fold Cross-Validation
    Data divided into k folds; model trained k times, each time using different fold as validation set
  29. 29Precision
    Of positive predictions, what fraction are correct? TP/(TP+FP); measures false positive rate
  30. 30Recall (Sensitivity)
    Of actual positives, what fraction are correctly identified? TP/(TP+FN); measures false negative rate

And 67 more cards. Study the whole deck free

Articles about this deck

More Technology decks

See all
Study all 97 cards free