Maths and Statistics for AI: The Essentials

Level 0, topic 3 of the BitByte AI path. A bilingual written tutorial with real, tested NumPy code; carefully chosen videos from 3Blue1Brown, StatQuest and Khan Academy; and three quizzes from easy to hard.

You start with what a vector is and finish able to explain how a model learns, how a language model picks its next word, and whether a reported improvement is real.

• Represent data as vectors and matrices and use dot products and cosine similarity (the heart of semantic search)
• Explain neural-network layers as W·x + b and read shapes in numeric code
• Explain derivatives, gradients, gradient descent, the learning rate and loss functions
• Summarise data with medians, spreads and percentiles, scale features, and tell correlation from causation
• Judge whether a difference is real with standard errors and bootstrap confidence intervals
• Use probability rules and Bayes' theorem, and explain temperature, top-k and top-p in LLMs
• Evaluate classifiers with a confusion matrix, precision, recall and F1, and choose a threshold

• Python basics, especially NumPy (the "Python for AI" course, Level 0 topic 2)
• School arithmetic: no calculus or statistics needed beforehand
• Basic English for the videos (the tutorial and quizzes are in English and Bangla)