The AI Taxonomy Tree - by Mohd Yamani Idna Idris
Artificial Intelligence (AI) ├── Symbolic AI (Rule-Based Systems) ├── Statistical AI │ └── Machine Learning (ML) │ ├── Supervised Learning │ ├── Unsupervised Learning │ └── Reinforcement Learning │ └── Deep Learning (DL) │ ├── MLP, CNN, RNN │ └── Transformers │ └── Large Language Models (LLMs) │ └── Generative AI │ ├── Text │ ├── Image │ ├── Audio │ └── Multimodal
AI Cheat Sheet: Algorithms, Types, Use Cases & Notes
| Algorithm / Model | Type | Use Case | Notes |
|---|---|---|---|
| Rule-Based Systems | Symbolic AI | Expert systems, diagnostics | No learning; logic-driven |
| Decision Trees | Supervised ML | Classification, regression | Easy to interpret |
| SVM | Supervised ML | Image/text classification | High-dimensional data |
| k-NN | Supervised ML | Pattern recognition | Simple, slow on large data |
| Linear Regression | Supervised ML | Predict continuous values | Assumes linearity |
| k-Means | Unsupervised ML | Customer segmentation | Requires k upfront |
| PCA | Unsupervised ML | Dimensionality reduction | Improves speed, loses detail |
| Q-Learning | Reinforcement Learning | Game AI, robotics | Learns via rewards |
| Policy Gradient | Reinforcement Learning | Continuous control | Direct policy optimization |
| MLP | Deep Learning | General prediction | Basic neural net |
| CNN | Deep Learning | Image recognition | Spatial feature extraction |
| RNN | Deep Learning | Sequence modeling | Struggles with long-term memory |
| LSTM / GRU | Deep Learning | Long-term dependencies | Better memory handling |
| Transformer | Deep Learning | NLP, vision, multimodal tasks | Scalable, parallelizable |
| BERT | Transformer (LLM) | Text classification, Q&A | Bidirectional context |
| GPT (e.g., GPT-4) | Transformer (LLM) | Text generation, chatbots | Autoregressive model |
| Stable Diffusion | Generative AI | Image synthesis | Noise-to-image pipeline |
| GAN | Generative AI | Image, video, music generation | Adversarial training |
| Jukebox / AudioLM | Generative AI | Music and audio creation | Trained on waveforms |
| CLIP | Multimodal AI | Text-to-image search | Connects vision + language |
| Gemini / GPT-4 Vision | Multimodal LLM | Image + text understanding | Combines modalities |
No comments:
Post a Comment