Level 1 - CETA Foundation
Course Curriculum
CETA Foundation – Basic ML
30+ 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗔𝗹𝗴𝗼𝗿𝗶𝘁𝗵𝗺
Ideal for freshers. Minimum +2 with decent English language skill.
1. 𝗦𝘂𝗽𝗲𝗿𝘃𝗶𝘀𝗲𝗱 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴
Predict with labeled data
Input (X) ──► Algorithm ──► Prediction (Y)
↳ Algorithms: 𝗟𝗶𝗻𝗲𝗮𝗿 𝗥𝗲𝗴𝗿𝗲𝘀𝘀𝗶𝗼𝗻, 𝗟𝗼𝗴𝗶𝘀𝘁𝗶𝗰 𝗥𝗲𝗴𝗿𝗲𝘀𝘀𝗶𝗼𝗻, 𝗗𝗲𝗰𝗶𝘀𝗶𝗼𝗻 𝗧𝗿𝗲𝗲, 𝗦𝗩𝗠, 𝗞-𝗡𝗡
2. 𝗨𝗻𝘀𝘂𝗽𝗲𝗿𝘃𝗶𝘀𝗲𝗱 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴
Find hidden patterns without labels
Raw Data ──► Clustering ──► Groups/Patterns
↳ Algorithms: 𝗞-𝗠𝗲𝗮𝗻𝘀, 𝗛𝗶𝗲𝗿𝗮𝗿𝗰𝗵𝗶𝗰𝗮𝗹, 𝗣𝗖𝗔
3. 𝗥𝗲𝗶𝗻𝗳𝗼𝗿𝗰𝗲𝗺𝗲𝗻𝘁 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴
Learn by reward & trial
[Agent] ──► Action ──► [Environment]
▲ │
└──── Reward ◄─────────┘
↳ Algorithms: 𝗤-𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴, 𝗗𝗲𝗲𝗽 𝗤-𝗡𝗲𝘁𝘄𝗼𝗿𝗸
4. 𝗘𝗻𝘀𝗲𝗺𝗯𝗹𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴
Combine weak models → stronger prediction
Model 1 ─┐
Mode 2 ─┼──► Voting/Boosting ──► Final Output
Model 3 ─┘
↳ Algorithms: 𝗥𝗮𝗻𝗱𝗼𝗺 𝗙𝗼𝗿𝗲𝘀𝘁, 𝗔𝗱𝗮𝗕𝗼𝗼𝘀𝘁, 𝗫𝗚𝗕𝗼𝗼𝘀𝘁
5. 𝗗𝗲𝗲𝗽 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴
Neural networks with layers
Input ──► [Hidden Layer 1] ──► [Hidden Layer 2] ──► Output
↳ Algorithms: 𝗖𝗡𝗡, 𝗥𝗡𝗡, 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗲𝗿𝘀
6. 𝗗𝗶𝗺𝗲𝗻𝘀𝗶𝗼𝗻𝗮𝗹𝗶𝘁𝘆 𝗥𝗲𝗱𝘂𝗰𝘁𝗶𝗼𝗻
Simplify high-dimensional data into fewer features
[100 Features] ──► PCA / t-SNE ──► [2D/3D Representation]
↳ Algorithms: 𝗣𝗖𝗔, 𝘁-𝗦𝗡𝗘, 𝗨𝗠𝗔𝗣
↳ Example: Reduce genomics data (10k genes) into few principal components for visualization
7. 𝗡𝗮𝘁𝘂𝗿𝗮𝗹 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗣𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴 (𝗡𝗟𝗣)
Understand and generate human language
Text ──► Tokenization ──► Embeddings ──► Model ──► Output
↳ Algorithms: 𝗕𝗮𝗴-𝗼𝗳-𝗪𝗼𝗿𝗱𝘀, 𝗪𝗼𝗿𝗱2𝗩𝗲𝗰, 𝗕𝗘𝗥𝗧, 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗲𝗿𝘀
↳ Example: Spam detection (Logistic Regression), ChatGPT-style Q&A (Transformers)
8. 𝗧𝗶𝗺𝗲 𝗦𝗲𝗿𝗶𝗲𝘀 & 𝗙𝗼𝗿𝗲𝗰𝗮𝘀𝘁𝗶𝗻𝗴
Predict future values from sequential data
Past Data ──► Model ──► Future Prediction
↳ Algorithms: 𝗔𝗥𝗜𝗠𝗔, SARIMA, 𝗟𝗦𝗧𝗠, 𝗣𝗿𝗼𝗽𝗵𝗲𝘁, BSTS
↳ Example: Stock price prediction, energy demand forecasting
9. 𝗔𝗻𝗼𝗺𝗮𝗹𝘆 𝗗𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻
Spot unusual data points in large sets
Normal Data Pattern ──► Model ──► Outlier Flag
↳ Algorithms: 𝗜𝘀𝗼𝗹𝗮𝘁𝗶𝗼𝗻 𝗙𝗼𝗿𝗲𝘀𝘁, 𝗟𝗼𝗰𝗮𝗹 𝗢𝘂𝘁𝗹𝗶𝗲𝗿, 𝗔𝘂𝘁𝗼𝗲𝗻𝗰𝗼𝗱𝗲𝗿
↳ Example: Credit card fraud detection, network intrusion detection
10. 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗠𝗼𝗱𝗲𝗹𝘀
Create new data that looks real
Noise ──► Generator ──► Synthetic Output
▲ │
└─── Discriminator ◄─┘
↳ Algorithms: 𝗚𝗔𝗡𝘀, 𝗩𝗔𝗘, 𝗗𝗶𝗳𝗳𝘂𝘀𝗶𝗼𝗻 𝗠𝗼𝗱𝗲𝗹𝘀
↳ Example: Deepfake images, AI art, synthetic medical data
11. Prompt Engg.
Transformers beyond language
Data ──► Preprocessing ──► Algorithm ──► Model ──► Prediction / Generation
↳ Open AI, Gemini, DeepSeek, Claud, Perplexity, Copilot, Grok, Fellou
12. C-in-C of AI
Publishing Open-AI Custom GPT – Building standalone systems using Prompts.
High Level Process
↳ Supplier ──► Input ──► Process ──► Output ──► Customer
Claude, Codex, Antigravity, Julius, Stitch, AIStudio, NotebookLm, n8n, ChatGPT, CustomGPts, Openclaw, nano-claw, Suno, Phot.ai, cover 171 AI platforms.