Artificial Intelligence Cheat Sheet
Complete Guide to AI, Machine Learning & Deep Learning Concepts
Core AI Concepts
Artificial Intelligence (AI)
Systems that can perform tasks requiring human-like intelligence—learning, reasoning, problem-solving, perception, and language understanding.
Machine Learning (ML)
Subset of AI where systems learn from data without explicit programming. The system improves performance through experience.
Deep Learning (DL)
Subset of ML using neural networks with multiple layers to learn hierarchical representations of data.
Neural Network
Computing system inspired by biological neural networks, consisting of interconnected nodes (neurons) that process information.
Types of Machine Learning
Supervised Learning
Training on labeled data (input-output pairs). Examples: classification, regression.
Unsupervised Learning
Finding patterns in unlabeled data. Examples: clustering, dimensionality reduction.
Reinforcement Learning
Learning through trial and error with rewards and penalties. Used in robotics, game playing, autonomous systems.
Semi-supervised Learning
Combines small amounts of labeled data with larger amounts of unlabeled data.
Common AI/ML Algorithms
- Linear/Logistic Regression: Predicting continuous values or binary outcomes
- Decision Trees: Tree-like models for classification and regression
- Random Forests: Ensemble of decision trees for improved accuracy
- Support Vector Machines (SVM): Finding optimal boundaries between classes
- K-Nearest Neighbors (KNN): Classification based on proximity to training examples
- K-Means Clustering: Grouping data into k clusters
- Gradient Boosting: Sequential ensemble method (XGBoost, LightGBM)
Neural Network Architectures
Convolutional Neural Networks (CNNs)
Specialized for image processing and computer vision.
Recurrent Neural Networks (RNNs)
Designed for sequential data like time series or text.
Long Short-Term Memory (LSTM)
Type of RNN that handles long-term dependencies.
Transformers
Architecture using attention mechanisms, foundation of modern LLMs (GPT, BERT).
Generative Adversarial Networks (GANs)
Two networks (generator and discriminator) competing to create realistic synthetic data.
Key AI Terms
Training Data
Dataset used to teach the model patterns.
Testing Data
Separate dataset used to evaluate model performance.
Validation Data
Data used during training to tune hyperparameters.
Overfitting
Model performs well on training data but poorly on new data (too complex).
Underfitting
Model is too simple to capture patterns in the data.
Bias
Error from oversimplified assumptions in the model.
Variance
Error from sensitivity to small fluctuations in training data.
Feature Engineering
Creating meaningful input variables from raw data.
Hyperparameters
Configuration settings chosen before training (learning rate, number of layers).
Epoch
One complete pass through the entire training dataset.
Batch Size
Number of training examples used in one iteration.
AI Applications
- Natural Language Processing (NLP): Text analysis, translation, chatbots, sentiment analysis
- Computer Vision: Image recognition, object detection, facial recognition
- Speech Recognition: Converting spoken language to text
- Recommendation Systems: Suggesting products, content, or connections
- Autonomous Vehicles: Self-driving cars and drones
- Robotics: Intelligent machines performing physical tasks
- Fraud Detection: Identifying unusual patterns in transactions
- Predictive Analytics: Forecasting future trends from historical data
Popular AI Frameworks & Tools
- TensorFlow: Google’s open-source ML platform
- PyTorch: Facebook’s deep learning framework
- Scikit-learn: Python library for classical ML algorithms
- Keras: High-level neural network API (now part of TensorFlow)
- Hugging Face: Platform for NLP models and transformers
- OpenCV: Computer vision library
Evaluation Metrics
Classification Metrics
Accuracy, Precision, Recall, F1-score, ROC-AUC
Regression Metrics
Mean Squared Error (MSE), Root Mean Squared Error (RMSE), R-squared
Confusion Matrix
Table showing true positives, false positives, true negatives, false negatives
Ethical Considerations
- Bias and fairness in AI systems
- Privacy and data protection
- Transparency and explainability
- Accountability for AI decisions
- Environmental impact of large models