| Both sides previous revisionPrevious revision | |
| ai [2026/01/22 09:31] – created phong2018 | ai [2026/02/22 23:54] (current) – removed phong2018 |
|---|
| ====== AI MODEL (Machine Learning & LLM) ====== | |
| |
| **What it is:** Artificial Intelligence models, đặc biệt là *Machine Learning* và *Large Language Models (LLMs)*, dùng thống kê và tối ưu số để học từ dữ liệu. | |
| |
| **What it’s for:** Tự động hoá suy luận, dự đoán, sinh nội dung và hỗ trợ ra quyết định trong hệ thống phần mềm. | |
| |
| --- | |
| |
| ===== Keyword Tree (click to open each child page) ===== | |
| |
| ==== 1) Foundation (Bắt buộc phải nắm) ==== | |
| |
| * [[ai:foundation:math-for-ai|Math for AI (Concept-level)]] | |
| |
| * [[ai:foundation:linear-algebra|Linear Algebra (Vector, Matrix)]] | |
| * [[ai:foundation:probability|Probability & Statistics]] | |
| * [[ai:foundation:gradient|Gradient & Optimization intuition]] | |
| * [[ai:foundation:ml-basics|Machine Learning Basics]] | |
| |
| * [[ai:foundation:supervised|Supervised Learning]] | |
| * [[ai:foundation:unsupervised|Unsupervised Learning]] | |
| * [[ai:foundation:reinforcement|Reinforcement Learning]] | |
| * [[ai:foundation:overfitting|Overfitting vs Underfitting]] | |
| |
| ==== 2) Deep Learning Core ==== | |
| |
| * [[ai:dl:neural-network|Neural Network]] | |
| |
| * [[ai:dl:neuron|Neuron & Weight]] | |
| * [[ai:dl:layer|Layer & Depth]] | |
| * [[ai:dl:activation|Activation Functions]] | |
| * [[ai:dl:training|Training Mechanics]] | |
| |
| * [[ai:dl:loss|Loss Function]] | |
| * [[ai:dl:backprop|Backpropagation (Concept)]] | |
| * [[ai:dl:gradient-descent|Gradient Descent]] | |
| |
| ==== 3) Transformer & LLM (Trái tim của AI hiện đại) ==== | |
| |
| * [[ai:llm:tokenization|Tokenization]] | |
| |
| * [[ai:llm:token-vs-word|Token vs Word]] | |
| * [[ai:llm:embedding|Embedding]] | |
| |
| * [[ai:llm:vector-space|Vector Space & Similarity]] | |
| * [[ai:llm:attention|Attention Mechanism]] | |
| |
| * [[ai:llm:self-attention|Self-Attention]] | |
| * [[ai:llm:qkv|Query – Key – Value]] | |
| * [[ai:llm:transformer|Transformer Architecture]] | |
| |
| * [[ai:llm:positional-encoding|Positional Encoding]] | |
| * [[ai:llm:multi-head|Multi-head Attention]] | |
| * [[ai:llm:decoder-only|Decoder-only (GPT-style)]] | |
| |
| ==== 4) LLM Training Pipeline ==== | |
| |
| * [[ai:training:pretraining|Pre-training]] | |
| * [[ai:training:finetuning|Fine-tuning]] | |
| * [[ai:training:instruction-tuning|Instruction Tuning]] | |
| * [[ai:training:rlhf|RLHF (Human Feedback)]] | |
| |
| ==== 5) Inference & Deployment ==== | |
| |
| * [[ai:inference:basics|Inference Basics]] | |
| |
| * [[ai:inference:training-vs-inference|Training vs Inference]] | |
| * [[ai:inference:latency|Latency & Throughput]] | |
| * [[ai:inference:context-window|Context Window]] | |
| * [[ai:inference:optimization|Model Optimization]] | |
| |
| * [[ai:inference:quantization|Quantization]] | |
| * [[ai:inference:distillation|Distillation]] | |
| |
| ==== 6) Prompting & Control ==== | |
| |
| * [[ai:prompting:prompt-structure|Prompt Structure]] | |
| |
| * [[ai:prompting:system-user-assistant|System / User / Assistant]] | |
| * [[ai:prompting:sampling|Sampling Parameters]] | |
| |
| * [[ai:prompting:temperature|Temperature]] | |
| * [[ai:prompting:top-p|Top-p]] | |
| * [[ai:prompting:risks|Prompt Risks]] | |
| |
| * [[ai:prompting:prompt-injection|Prompt Injection]] | |
| |
| ==== 7) AI + Data (Production AI) ==== | |
| |
| * [[ai:data:embedding-search|Embedding Search]] | |
| |
| * [[ai:data:vector-db|Vector Database]] | |
| * [[ai:data:rag|RAG (Retrieval Augmented Generation)]] | |
| * [[ai:data:tool-calling|Tool / Function Calling]] | |
| * [[ai:data:agent|Agent Pattern]] | |
| |
| ==== 8) Limitations & Failure Modes ==== | |
| |
| * [[ai:limits:hallucination|Hallucination]] | |
| * [[ai:limits:non-determinism|Non-determinism]] | |
| * [[ai:limits:security|Security & Data Leakage]] | |
| |
| ==== 9) Evaluation & Governance ==== | |
| |
| * [[ai:governance:evaluation|AI Evaluation]] | |
| |
| * [[ai:governance:golden-dataset|Golden Dataset]] | |
| * [[ai:governance:regression-test|AI Regression Test]] | |
| * [[ai:governance:human-in-loop|Human-in-the-loop]] | |
| * [[ai:governance:ai-sdlc|AI-SDLC]] | |
| |
| ==== 10) Mastery Check ==== | |
| |
| * [[ai:mastery:why-model-fails|Explain why model fails]] | |
| * [[ai:mastery:when-not-to-use-ai|When NOT to use AI]] | |
| * [[ai:mastery:ai-resilient-system|Design AI-resilient systems]] | |
| |
| |
| |
