AI/ML models
AI/ML models vary from basic reactive systems to complex deep learning, categorized broadly by learning style (Supervised, Unsupervised, Reinforcement) or capability (Narrow, General), with specific algorithms like Linear Regression, Decision Trees, SVMs, Neural Networks, and Transformers handling tasks from classification and prediction to image/language generation.
By Learning Style
- Supervised Learning: Learns from labeled data to make predictions (e.g., predicting house prices, classifying emails).
- Examples: Linear Regression, Logistic Regression, Decision Trees, Random Forest, SVM, Naive Bayes.
- Unsupervised Learning: Finds patterns in unlabeled data (e.g., customer segmentation).
- Reinforcement Learning (RL): Learns through trial-and-error with rewards/penalties (e.g., game AI, robotics).
- Semi-Supervised/Self-Supervised: Uses a mix of labeled/unlabeled data or learns from data structure itself.
By Capability/Type
- Narrow AI (Weak AI): Performs specific tasks (e.g., Siri, Google Search).
- General AI (Strong AI): Human-level intelligence (theoretical).
- Deep Learning Models: Use neural networks with many layers for complex tasks like computer vision & NLP.
- Examples: Convolutional Neural Networks (CNNs) for images, Recurrent Neural Networks (RNNs) for sequences, Transformers for language (like GPT).
- Rule-Based/Expert Systems: Older AI using predefined rules.
Common Algorithms & Models
- Regression: Predicts continuous values (Linear, Logistic).
- Classification: Categorizes data (Naive Bayes, SVM, Decision Trees, KNN).
- Clustering: Groups similar data points (e.g., K-Means, used in Unsupervised Learning).
- Ensemble Methods: Combine multiple models (Random Forest, Gradient Boosting).
- Transformers: Revolutionized NLP and vision (e.g., GPT, BERT).
Modern Applications
- Language Models (LLMs): GPT, Claude, Gemini (Chat, Text Generation).
- Image/Video Models: Veo, Sora (Image/Video Generation).
- Computer Vision Models: Facial Recognition, Object Detection.