Skip to content

README.md

Latest commit

 

History

History
229 lines (171 loc) · 7.59 KB

README.md

File metadata and controls

229 lines (171 loc) · 7.59 KB

CMU MSML Preparation - Practice Problems & Overviews

Purpose: Organized collection of ML fundamentals with CMU MSML course alignment and research connections.

Structure:

  • 00_Skill_Exercises/ - ⭐ START HERE! Completed solutions for skill-building exercises
  • Templates/ - ⭐ PRACTICE HERE! Blank templates to practice before looking at solutions
  • 01_Foundations/ - Core concepts (ReLU, Gradient Descent, Matrix Ops)
  • 02_Course_Specific/ - Problems organized by CMU courses
  • 03_Practice_Problems/ - Additional practice with solutions
  • 04_Research_Alignment/ - Connections to CMU research (Dr. Shah, etc.)

🚀 Quick Start Guide

For Complete Beginners:

  1. Start with Templates/ - Practice with blank templates first
    • Begin with Templates/01_Core_Python/exercise_1_normalize_SIMPLE.py (easiest!)
    • Read Templates/01_Core_Python/SIMPLE_GUIDE.md for step-by-step help
    • Use Templates/HOW_TO_SOLVE.md for problem-solving strategies
  2. Compare with solutions in 00_Skill_Exercises/ after attempting
  3. Use side-by-side guides like Templates/02_NumPy/SIDE_BY_SIDE_COMPARISON.md

For CMU Prep:

  1. Master fundamentals in 00_Skill_Exercises/ and Templates/
  2. Review 01_Foundations/ - Core concepts aligned with CMU courses
  3. Review 02_Course_Specific/ by course number
  4. Focus on courses you'll take first semester
  5. Connect problems to research interests in 04_Research_Alignment/

For Practice:

  1. Use Templates/ for blank practice exercises
  2. Try problems without looking at solutions first
  3. Compare your approach with solutions in 00_Skill_Exercises/
  4. Use 03_Practice_Problems/ for additional exercises

CMU Course Mapping

10-701/715: Introduction to Machine Learning

  • Evaluation Metrics (Accuracy, Precision, Recall)
  • Linear Regression (Gradient Descent, Normal Equation)
  • Feature Scaling
  • One-Hot Encoding

10-617/707: Deep Learning

  • ReLU Activation Function
  • Softmax Activation Function
  • Matrix Operations (foundation)
  • Neural Network Fundamentals

10-718: Machine Learning in Practice

  • Feature Scaling
  • Evaluation Metrics
  • One-Hot Encoding
  • Preprocessing Pipeline

10-725: Optimization for Machine Learning

  • Gradient Descent
  • Matrix Operations
  • Linear Regression
  • Optimization Fundamentals

36-700/705: Probability & Statistics

  • Matrix Operations (foundation)
  • Evaluation Metrics (statistical perspective)

Research Alignment

Dr. Shah's Research Areas:

  • Evaluation Science
  • Annotation Bias
  • Reviewer Assignment
  • "The More You Automate, The Less You See"

Relevant Problems:

  • Evaluation Metrics (Calculate Accuracy Score)
  • Feature Engineering (One-Hot Encoding)
  • Preprocessing (Feature Scaling)

Problem Difficulty Levels

  • ⭐ Beginner: Start here! Builds fundamentals
  • ⭐⭐ Intermediate: Requires understanding of basics
  • ⭐⭐⭐ Advanced: For deeper understanding

📁 Folder Structure Details

00_Skill_Exercises/ - Completed Solutions

Purpose: Reference solutions for skill-building exercises

Available Exercises:

  • 01_Core_Python/exercise_1_normalize.py - Z-score normalization
  • 02_NumPy/exercise_2_matrix_ops.py - Matrix operations & broadcasting
  • 03_Loss_Gradients/exercise_3_mse_loss.py - Loss functions & gradients

How to Use:

  • ⚠️ Don't peek until you've tried! Use Templates/ first
  • Compare your implementation after attempting
  • Learn from differences in approach

Templates/ - Practice Templates

Purpose: Blank templates for practicing before looking at solutions

Available Templates:

  • 01_Core_Python/
    • exercise_1_normalize_SIMPLE.py - ⭐ EASIEST VERSION! Simplified with step-by-step hints
    • exercise_1_normalize_template.py - Standard template
    • SIMPLE_GUIDE.md - ⭐ START HERE! Step-by-step problem-solving guide
  • 02_NumPy/
    • exercise_2_matrix_ops_template.py - Matrix operations practice
    • SIDE_BY_SIDE_COMPARISON.md - ⭐ NEW! Template vs solution comparison
    • NEXT_STEPS.md - Guidance for next exercises
  • 03_Loss_Gradients/
    • exercise_3_mse_loss_template.py - Loss function implementation

Helpful Guides:

  • HOW_TO_SOLVE.md - General problem-solving strategies
  • WALKTHROUGH_EXAMPLE.md - Complete detailed example walkthrough
  • README.md - Template usage instructions

How to Use:

  1. Open a template file
  2. Fill in TODO sections
  3. Test your implementation
  4. Compare with 00_Skill_Exercises/ solutions

01_Foundations/ - Core ML Concepts

Purpose: Fundamental ML concepts aligned with CMU courses

Topics:

  • Activation_Functions/ - ReLU, Softmax, Sigmoid
  • Gradient_Descent/ - Linear regression optimization
  • Matrix_Operations/ - Matrix-vector multiplication, reshaping
  • Preprocessing/ - Feature scaling, one-hot encoding
  • Evaluation/ - Accuracy, precision, recall metrics

02_Course_Specific/ - CMU Course Alignment

Purpose: Problems organized by specific CMU courses

Courses Covered:

  • 10-701/715: Introduction to Machine Learning
  • 10-617/707: Deep Learning
  • 10-718: Machine Learning in Practice
  • 10-725: Optimization for Machine Learning
  • 36-700/705: Probability & Statistics

03_Practice_Problems/ - Additional Practice

Purpose: Extra exercises for reinforcement

04_Research_Alignment/ - Research Connections

Purpose: Connect problems to CMU research areas (Dr. Shah, etc.)

🎯 Recommended Learning Path

Week 1-2: Core Python Fundamentals

  1. Start with Templates/01_Core_Python/exercise_1_normalize_SIMPLE.py
  2. Read Templates/01_Core_Python/SIMPLE_GUIDE.md
  3. Complete the normalization exercise
  4. Compare with 00_Skill_Exercises/01_Core_Python/exercise_1_normalize.py

Week 3-4: NumPy & Matrix Operations

  1. Work through Templates/02_NumPy/exercise_2_matrix_ops_template.py
  2. Use Templates/02_NumPy/SIDE_BY_SIDE_COMPARISON.md for reference
  3. Master matrix multiplication, broadcasting, vectorized operations
  4. Compare with 00_Skill_Exercises/02_NumPy/exercise_2_matrix_ops.py

Week 5-6: Loss Functions & Gradients

  1. Practice with Templates/03_Loss_Gradients/exercise_3_mse_loss_template.py
  2. Understand MSE, MAE, and gradient computation
  3. Compare with 00_Skill_Exercises/03_Loss_Gradients/exercise_3_mse_loss.py

Week 7+: Foundations & Course-Specific

  1. Review 01_Foundations/ concepts
  2. Work through 02_Course_Specific/ problems
  3. Explore 04_Research_Alignment/ connections

How to Use This Folder

  1. Practice Mode: Use Templates/ - Implement solutions yourself
  2. Study Mode: Read 01_Foundations/ - Understand concepts
  3. Review Mode: Compare with 00_Skill_Exercises/ - Learn from solutions
  4. Research Mode: Explore 04_Research_Alignment/ - Connect to CMU research

📚 Additional Resources

  • QUICK_START.md - Quick reference guide
  • ENHANCED_WORKFLOW.md - Complete 12-16 week learning path
  • SKILLS_CHECKLIST.md - Track your progress
  • STUDY_PLAN.md - Structured study schedule
  • MASTER_INDEX.md - Complete index of all problems

Next Steps

  1. ✅ Start with Templates - Begin with Templates/01_Core_Python/exercise_1_normalize_SIMPLE.py
  2. ✅ Master fundamentals - Complete all exercises in 00_Skill_Exercises/
  3. ✅ Review foundations - Work through 01_Foundations/
  4. ✅ Course prep - Review 02_Course_Specific/ problems
  5. ✅ Practice - Use 03_Practice_Problems/ for reinforcement
  6. ✅ Research - Explore 04_Research_Alignment/ connections

Good luck with your CMU MSML preparation! 🎓