Wiki.Quizz.vn

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== 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]]