AI & Machine Learning
Concepts, models, tools, and prompt engineering — referenced and explained clearly.
Foundational Concepts
Machine Learning Types
- Supervised learning — trained on labeled data (classification, regression)
- Unsupervised learning — finding patterns in unlabeled data (clustering, dimensionality reduction)
- Reinforcement learning — agent learns via reward/penalty feedback
Key Terms
| Term | Definition |
|---|---|
| Model | A mathematical function trained on data to make predictions |
| Training | Adjusting model weights to minimize loss on a dataset |
| Inference | Running a trained model to produce output |
| Fine-tuning | Further training a pre-trained model on specific data |
| Embeddings | Dense vector representations of data (text, images, etc.) |
| Tokens | Units of text processed by language models |
Large Language Models (LLMs)
LLMs are transformer-based models trained on vast text corpora to predict and generate text.
Prompt Engineering Principles
- Be specific — clear instructions outperform vague ones
- Provide context — include relevant background
- Use examples — few-shot prompting improves output quality
- Define the format — ask for JSON, bullet points, tables as needed
- Iterate — treat prompts as code: version and refine
Common Patterns
- RAG (Retrieval-Augmented Generation) — ground LLM responses in retrieved documents
- Chain of Thought — prompt the model to reason step by step
- Tool use / Function calling — LLM selects and calls predefined functions
References
- Goodfellow, I., Bengio, Y., & Courville, A. Deep Learning. MIT Press, 2016. https://www.deeplearningbook.org
- Anthropic. Claude Documentation. https://docs.anthropic.com
- OpenAI. Prompt Engineering Guide. https://platform.openai.com/docs/guides/prompt-engineering
Last updated on