Source Library

This page is provenance for the solved-question layer.

It is not the first place to start. Solve a pack first, notice the weak concept, and only then use this page to understand where that style of question came from.

How To Use This Page

Use this page in a short loop:

1. solve a pack from Solved Questions
2. notice the weak theme
3. find the source family that matches that theme
4. return to the matching academy topic or track

The point is not to send the learner into a pile of old PDFs. The point is to keep provenance clear while the academy stays pedagogically focused.

What The Academy Is Doing With These Sources

The academy questions are:

• based on public official materials
• adapted into academy wording
• shortened or restructured to fit the topic and workflow style here

They are not verbatim copies of original exams or handouts.

Source Families

MIT OpenCourseWare 6.867

• 6.867 Machine Learning course page
• 6.867 exam solutions archive

Good for:

• SVM margins and feature maps
• generative versus discriminative comparisons
• active-learning and logistic-regression intuition

Best academy follow-up:

• SVM, Kernels, and Learning Theory
• SVM Margins and Kernels

Carnegie Mellon 10-601

• 10-601 previous materials and exam archive
• Spring 2009 midterm solutions

Good for:

• Naive Bayes smoothing and priors
• regression model comparison
• decision trees versus linear models
• learning-theory style reasoning

Best academy follow-up:

• Validation, Leakage, and Model Choice
• Honest Splits and Baselines

Stanford CS229

• CS229 Spring 2019 schedule and materials
• CS229 Problem Set 1, Summer 2020

Good for:

• feature maps
• logistic versus generative modeling choices
• model selection and bias-variance reasoning
• k-means, PCA, and unsupervised-learning review

Best academy follow-up:

• Unsupervised Learning and Representation
• Clustering and Low-Dimensional Views

Stanford CS231n

• CS231n schedule
• CS231n assignments

Good for:

• optimization and learning-rate reasoning
• dropout, batch normalization, and regularization choices
• checkpoint selection
• transfer-learning decisions

Best academy follow-up:

• Deep Learning and Checkpoints
• PyTorch Training Loops

UC Berkeley CS189

• CS189 past exams
• CS189 Fall 2017 archive

Good for:

• OLS and normal-equation reasoning
• computational cost and gradient-descent updates
• ridge, cross-validation, and PCA themes
• bias-variance and evaluation questions

Best academy follow-up:

• Validation, Leakage, and Model Choice
• Cross-Validation

Cornell CS4786

• CS4786 placement exam
• CS4786 practice prelim page

Good for:

• optimization with constraints
• principal-eigenvector and Rayleigh-quotient reasoning
• placement-style linear algebra drills

Best academy follow-up:

• Theory Checkpoint Sheets
• Deep Learning and Checkpoints

What This Page Is Not

This page is not:

• a reading assignment before you solve questions
• a mirror of university archives
• a substitute for academy topics, clinics, or tracks

Use it to keep the question layer honest and sourced, then go back to the active learning loop.