{
 "source": "https://peabody.kids/flashcards/tech-ai-machine-learning-terms",
 "id": "tech-ai-machine-learning-terms",
 "title": "AI and Machine Learning Terms",
 "category": "technology",
 "subject": "Artificial Intelligence",
 "levels": [
  "6-8",
  "9-12",
  "college",
  "adult"
 ],
 "description": "Key ideas in artificial intelligence, from training data to neural networks.",
 "tags": [
  "AI",
  "machine learning",
  "neural networks",
  "LLM"
 ],
 "frontLabel": "Term",
 "backLabel": "Meaning",
 "cards": [
  [
   "Artificial intelligence (AI)",
   "Computer systems that do tasks that usually need human intelligence"
  ],
  [
   "Machine learning (ML)",
   "Computers learning patterns from data instead of following only fixed rules"
  ],
  [
   "Deep learning",
   "Machine learning with many-layered neural networks"
  ],
  [
   "Neural network",
   "A system of connected nodes inspired loosely by brain cells"
  ],
  [
   "Neuron (node)",
   "A unit in a network that takes inputs and gives an output"
  ],
  [
   "Layer",
   "A group of nodes in a neural network"
  ],
  [
   "Weights",
   "Numbers in a network that are adjusted during learning"
  ],
  [
   "Parameters",
   "The values a model learns, such as weights"
  ],
  [
   "Training",
   "Teaching a model using data"
  ],
  [
   "Training data",
   "Examples the model learns from"
  ],
  [
   "Test data",
   "New examples used to check how well a model works"
  ],
  [
   "Validation data",
   "Data used to tune a model during development"
  ],
  [
   "Label",
   "The correct answer attached to an example"
  ],
  [
   "Feature",
   "A measurable property used as input"
  ],
  [
   "Model",
   "The result of training; it makes predictions"
  ],
  [
   "Prediction (inference)",
   "The model's output on new data"
  ],
  [
   "Supervised learning",
   "Learning from labeled examples"
  ],
  [
   "Unsupervised learning",
   "Finding patterns in unlabeled data"
  ],
  [
   "Reinforcement learning",
   "Learning by trial and error with rewards"
  ],
  [
   "Classification",
   "Sorting into categories, like cat or dog"
  ],
  [
   "Regression",
   "Predicting a number, like a house price"
  ],
  [
   "Clustering",
   "Grouping similar items together"
  ],
  [
   "Overfitting",
   "A model that memorizes training data and does poorly on new data"
  ],
  [
   "Underfitting",
   "A model too simple to capture patterns"
  ],
  [
   "Bias (in AI)",
   "Unfair results caused by skewed data or design"
  ],
  [
   "Fairness",
   "Making AI treat people equitably"
  ],
  [
   "Algorithm",
   "Step-by-step instructions"
  ],
  [
   "Gradient descent",
   "A method for adjusting weights to reduce errors"
  ],
  [
   "Loss function",
   "A measure of how wrong the model is"
  ],
  [
   "Epoch",
   "One full pass through the training data"
  ],
  [
   "Learning rate",
   "How big the steps are when updating weights"
  ],
  [
   "Accuracy",
   "The share of predictions that are correct"
  ],
  [
   "Confusion matrix",
   "A table of correct and wrong predictions by class"
  ],
  [
   "Computer vision",
   "AI that understands images and video"
  ],
  [
   "Natural language processing (NLP)",
   "AI that works with human language"
  ],
  [
   "Speech recognition",
   "Turning speech into text"
  ],
  [
   "Text-to-speech",
   "Turning text into spoken sound"
  ],
  [
   "Chatbot",
   "A program that converses with people"
  ],
  [
   "Large language model (LLM)",
   "A very large model trained on lots of text to predict and generate language"
  ],
  [
   "Token",
   "A piece of text, such as a word or part of a word, that language models process"
  ],
  [
   "Prompt",
   "The input or instruction given to an AI model"
  ],
  [
   "Prompt engineering",
   "Writing prompts to get better answers"
  ],
  [
   "Context window",
   "How much text a model can consider at once"
  ],
  [
   "Transformer",
   "A neural network design behind most modern language models"
  ],
  [
   "Attention",
   "A mechanism that lets a model focus on relevant parts of its input"
  ],
  [
   "Generative AI",
   "AI that creates new text, images, audio or code"
  ],
  [
   "Image generation",
   "Creating pictures from text descriptions"
  ],
  [
   "Diffusion model",
   "A model that creates images by gradually removing noise"
  ],
  [
   "Hallucination",
   "When an AI states something false as if it were true"
  ],
  [
   "Fine-tuning",
   "Further training a model on specialized data"
  ],
  [
   "Pre-training",
   "Initial training on a large general dataset"
  ],
  [
   "Embedding",
   "A list of numbers representing the meaning of text or images"
  ],
  [
   "Retrieval-augmented generation (RAG)",
   "Letting a model look up documents before answering"
  ],
  [
   "AI agent",
   "An AI system that can take actions toward a goal"
  ],
  [
   "Turing test",
   "Alan Turing's idea of judging a machine by whether it can seem human in conversation"
  ],
  [
   "Alan Turing",
   "British mathematician who asked 'Can machines think?' in 1950"
  ],
  [
   "Dartmouth workshop",
   "The 1956 meeting where the term artificial intelligence was coined"
  ],
  [
   "Deep Blue",
   "IBM computer that beat chess champion Garry Kasparov in 1997"
  ],
  [
   "AlphaGo",
   "DeepMind program that beat Go champion Lee Sedol in 2016"
  ],
  [
   "ImageNet",
   "A huge image dataset that sparked deep learning progress"
  ],
  [
   "Narrow AI",
   "AI built for specific tasks"
  ],
  [
   "General AI (AGI)",
   "Hypothetical AI that can do any intellectual task a human can"
  ],
  [
   "Robotics",
   "Building machines that sense and act"
  ],
  [
   "Self-driving car",
   "A vehicle that drives using sensors and AI"
  ],
  [
   "Recommendation system",
   "AI that suggests videos, songs or products"
  ],
  [
   "Deepfake",
   "A fake image or video made by AI"
  ],
  [
   "AI ethics",
   "Thinking about fairness, safety, privacy and responsibility in AI"
  ],
  [
   "Privacy and AI",
   "AI systems should protect people's personal information"
  ],
  [
   "Human in the loop",
   "A person reviews or guides the AI's decisions"
  ],
  [
   "Explainability",
   "Understanding why an AI made a decision"
  ],
  [
   "Check AI answers",
   "AI can be wrong, so verify important facts"
  ]
 ]
}