By Capability
-
Narrow AI (Weak AI)
Examples: Siri, ChatGPT, Netflix recommendations
-
General AI (AGI)
Still theoretical: Can perform any intellectual task like a human
-
Superintelligent AI
Hypothetical future AI that surpasses human intelligence
By Functionality
-
Reactive Machines
No memory; only responds to inputs. Example: Deep Blue (chess)
-
Limited Memory
Learns from historical data. Example: Self-driving cars
Self-driving AI cars have limited memory. Will this improve? Answer
-
Theory of Mind
Hypothetical; understands emotions, beliefs more
-
Self-aware AI
Future concept; AI with consciousness
By Technology
-
Machine Learning
Supervised Learning, Unsupervised Learning, Reinforcement Learning, Self-Supervised Learning more
-
Deep Learning
Uses neural networks (e.g., CNNs, RNNs, Transformers)
-
Natural Language Processing (NLP)
Sentiment analysis, text generation, speech recognition
-
Computer Vision
Image recognition, object detection, facial recognition
-
Expert Systems
Rule-based decision making
-
Robotics
Autonomous machines using AI (e.g., Boston Dynamics robots)
-
Evolutionary Algorithms
Optimization based on natural selection principles
-
Symbolic AI
Logic and rules-based systems
AI Concept Map
Many of these sub-links to be explored at a later date
Generative AI
ChatGPT said: The diagram you shared is a concept map of Artificial Intelligence (AI) and its major subfields. However, it doesn't explicitly include Generative AI, which is a rapidly growing and impactful area of AI today.
Where Does Generative AI Fit In?
Generative AI refers to AI systems that can create new content — such as text, images, audio, or code — often using deep learning models. It fits under Machine Learning and more specifically under Deep Learning.
Here's how it could be added to the diagram:
- Artificial Intelligence
- Machine Learning
- Deep Learning
- Generative AI
- Text Generation (e.g., GPT)
- Image Generation (e.g., DALL·E) example
- Music, Video, Code, etc.
Generative AI often utilizes transformer-based models, GANs (Generative Adversarial Networks), or diffusion models, depending on the application.
Transformer-based models and diffusion models here.
GANs (Generative Adversarial Networks) here.
We won't ask ChatGPT to add that lot to an imperfect concept map.