LNET/lnet_tutor
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
8dc869634e5997449c27ec47dce4fdd0eea45b22
lnet_tutor/learning_plans/linear_algebra/README.md
howard 8dc869634e first working version
2025-10-22 20:14:31 +08:00

10 KiB
Raw Blame History

Linear Algebra Learning Plan

📐 Welcome to Your Linear Algebra Mastery Journey!

This comprehensive learning plan will guide you from basic vectors to advanced applications in machine learning, computer graphics, and data science.

📚 What's Included

1. Master Plan (00_LINEAR_ALGEBRA_MASTER_PLAN.md)

Your complete roadmap containing:

  • 22 detailed modules organized in 5 phases
  • From geometric intuition to abstract theory
  • Applications in ML, graphics, data science
  • Resource recommendations (textbooks, videos, tools)
  • Milestone achievements with project ideas
  • Specialization paths (ML, Graphics, Quantum, Computational)

2. Knowledge Graph (01_KNOWLEDGE_GRAPH.md)

Complete dependency map showing:

  • 15 knowledge levels from basics to expert
  • Topic dependencies clearly mapped
  • Parallel learning opportunities
  • Visual knowledge tree
  • Critical learning path

3. Initial Assessment (02_INITIAL_ASSESSMENT.md)

Determine your starting point with:

  • Self-assessment covering 40+ topics
  • 6 computational problems (beginner to expert)
  • Proficiency level determination
  • Personalized recommendations

4. Assessments Directory (assessments/)

Track your exam performance:

  • Personalized assessments after each exam
  • Strengths and weaknesses identified
  • Progress tracking over time

🎯 Learning Path Overview

Phase 1: Foundations (1.5-2 months)

Goal: Master vectors and matrices

  • Module 1.1: Vectors Basics (geometric)
  • Module 1.2: Dot Product & Vector Operations
  • Module 1.3: Matrices Basics
  • Module 1.4: Matrix Properties

Phase 2: Core Theory (2-3 months)

Goal: Master systems, decompositions, eigenvalues

  • Module 2.1: Systems of Linear Equations
  • Module 2.2: Matrix Inverses
  • Module 2.3: Determinants
  • Module 2.4: Vector Spaces
  • Module 2.5: Linear Transformations
  • Module 2.6: Eigenvalues & Eigenvectors

Phase 3: Advanced Topics (1.5-2 months)

Goal: Master orthogonality and decompositions

  • Module 3.1: Orthogonality
  • Module 3.2: Inner Product Spaces
  • Module 3.3: Matrix Decompositions (LU, QR, SVD)
  • Module 3.4: Norms & Conditioning

Phase 4: Applications (1-2 months)

Goal: Apply to real-world problems

  • Module 4.1: Machine Learning (PCA, regression)
  • Module 4.2: Computer Graphics (transformations)
  • Module 4.3: Optimization
  • Module 4.4: Data Science

Phase 5: Specialization (Ongoing)

Choose your path:

  • Machine Learning Deep Dive
  • Computational Linear Algebra
  • Quantum Computing
  • Advanced Applications

🚀 Quick Start

Step 1: Prerequisites (Optional, 1-2 days)

  • Review basic algebra if rusty
  • Set up Python + NumPy OR MATLAB
  • Test with simple calculations

Step 2: Assessment (1-2 hours)

  1. Open 02_INITIAL_ASSESSMENT.md
  2. Complete self-assessment
  3. Try computational problems
  4. Determine your level

Step 3: Build Intuition (1 week)

  1. WATCH: 3Blue1Brown "Essence of Linear Algebra" (11 videos, ~3 hours total)
  2. This series provides incredible geometric intuition
  3. Watch before heavy studying!

Step 4: Study (Daily)

  1. Read theory (30-40 min)
  2. Solve problems (30-40 min)
  3. Prove theorems (20-30 min)
  4. Code implementations (optional)

💻 Recommended Tools

Python + NumPy (Recommended for Programmers)

import numpy as np

# Vectors
v = np.array([1, 2, 3])
w = np.array([4, 5, 6])
dot = np.dot(v, w)          # Dot product
norm = np.linalg.norm(v)    # Magnitude

# Matrices
A = np.array([[1, 2], [3, 4]])
B = np.linalg.inv(A)        # Inverse
det = np.linalg.det(A)      # Determinant
eig = np.linalg.eig(A)      # Eigenvalues

# Solve systems
x = np.linalg.solve(A, b)   # Solve Ax = b

# Decompositions
U, S, Vt = np.linalg.svd(A) # SVD
Q, R = np.linalg.qr(A)      # QR

MATLAB/Octave (Industry Standard)

% Matrices are first-class citizens
A = [1 2; 3 4];
B = inv(A);              % Inverse
det_A = det(A);          % Determinant
[V, D] = eig(A);         % Eigenvalues

% Solve systems
x = A \ b;               % Solve Ax = b

% Decompositions
[U, S, V] = svd(A);      % SVD
[Q, R] = qr(A);          % QR

📚 Essential Resources

Must-Watch Videos

  1. 3Blue1Brown: "Essence of Linear Algebra" (11 videos)
    • BEST visual intuition
    • Watch FIRST before anything else
    • Free on YouTube

Textbooks (In Order)

  1. "Introduction to Linear Algebra" by Gilbert Strang

    • Best overall introduction
    • Clear explanations
    • Many applications

  2. "Linear Algebra and Its Applications" by David Lay

    • Very accessible
    • Application-focused
    • Great for beginners

  3. "Linear Algebra Done Right" by Sheldon Axler

    • More theoretical
    • Avoids determinants initially
    • Beautiful proofs

  4. "Matrix Analysis" by Horn & Johnson

    • Advanced reference
    • Comprehensive
    • For deep study

Online Courses

  • MIT OCW: Gilbert Strang's 18.06 (legendary!)
  • Khan Academy: Linear Algebra series
  • Brilliant.org: Interactive problems

🏆 Key Milestones

Milestone 1: Vector & Matrix Fluency ✅

  • Timing: Month 2
  • Skills: All vector/matrix operations
  • Project: Vector/matrix library in Python
  • Test: Solve 20 problems in 30 minutes

Milestone 2: Systems Mastery ✅

  • Timing: Month 4-5
  • Skills: Solve any linear system, compute inverses
  • Project: Linear equation solver
  • Test: Pass comprehensive exam (75%+)

Milestone 3: Eigenvalue Mastery ✅

  • Timing: Month 6-7
  • Skills: Eigenvalues, eigenvectors, diagonalization
  • Project: Markov chain simulator
  • Test: Pass advanced exam (70%+)

Milestone 4: SVD & Applications ✅

  • Timing: Month 8-9
  • Skills: SVD, PCA, graphics transforms
  • Project: Image compression or PCA implementation
  • Test: Apply to real data

Milestone 5: Specialization ✅

  • Timing: Month 10+
  • Skills: Deep expertise in chosen area
  • Project: ML model, graphics engine, or quantum algorithm
  • Certification: Professional portfolio

💡 Linear Algebra Learning Tips

Do's ✅

  • Visualize everything - Draw vectors and transformations
  • Use 3Blue1Brown - Best intuition builder
  • Solve many problems - Fluency requires practice
  • Implement in code - Programming solidifies understanding
  • Prove key theorems - Understand WHY, not just HOW
  • Connect to applications - See real-world relevance
  • Start geometric - Intuition before abstraction

Don'ts ❌

  • Don't memorize formulas without understanding
  • Don't skip geometric interpretation
  • Don't avoid proofs entirely
  • Don't neglect computational practice
  • Don't rush through fundamentals
  • Don't study in isolation (use visualizations)

🎯 Why Learn Linear Algebra?

Foundation for Modern Tech

  • Machine Learning: PCA, neural networks, optimization
  • Computer Graphics: ALL transformations are matrices
  • Data Science: Dimensionality reduction, analysis
  • Quantum Computing: Quantum states are vectors
  • Computer Vision: Image processing, feature extraction
  • Natural Language Processing: Word embeddings, transformers

Real Applications

  • Netflix recommendations (SVD, matrix factorization)
  • Google PageRank (eigenvectors of web graph)
  • Face recognition (eigenfaces, PCA)
  • 3D video games (transformation matrices)
  • Self-driving cars (sensor fusion, optimization)
  • ChatGPT/LLMs (attention is matrix operations!)

Career Impact

  • Required for ML engineer roles
  • Essential for data science
  • Critical for graphics programming
  • Foundation for AI research
  • Needed for quantitative finance

📊 Study Schedules

Full-Time (3-4 hours/day)

  • Timeline: 5-6 months to applications
  • Daily: 1 hour theory + 1-2 hours problems + 1 hour coding
  • Projects: 1-2 per week
  • Pace: 1 module per week

Part-Time (1.5-2 hours/day)

  • Timeline: 8-10 months to applications
  • Daily: 40 min theory + 40 min problems + 20 min review
  • Projects: 1 per week
  • Pace: 1 module per 1.5-2 weeks

Casual (1 hour/day)

  • Timeline: 12-15 months to applications
  • Daily: 30 min theory + 30 min problems
  • Projects: 2 per month
  • Pace: 1 module per 2-3 weeks

🎓 Integration with Tech Learning

Python Integration

Use NumPy to implement all concepts:

  • Vectors and matrices
  • Linear transformations
  • Eigenvalue computation
  • SVD and PCA
  • ML applications

C++ Integration

Implement for performance:

  • Matrix libraries
  • Graphics transformations
  • Game engine math
  • Scientific computing

Machine Learning

Linear algebra is EVERYWHERE:

  • Data representation
  • Model parameters
  • Forward/backward pass
  • Optimization
  • Dimensionality reduction

🌟 What Makes This Plan Special

Visual & Intuitive

  • Emphasizes geometric understanding
  • 3Blue1Brown integration
  • Visualization tools
  • Draw everything!

Computation & Theory Balanced

  • 60% computational practice
  • 25% theoretical understanding
  • 15% applications
  • Learn by doing AND understanding

Application-Driven

  • See real uses immediately
  • Build actual projects
  • Connect to ML, graphics, data science
  • Not just abstract math

Modern & Practical

  • Python/NumPy focus
  • Industry-relevant skills
  • Modern applications (ML, AI)
  • Cutting-edge topics

🎯 Your Next Steps

  1. ☐ Read this README
  2. ☐ WATCH: 3Blue1Brown videos 1-3 (build intuition!)
  3. ☐ Complete 02_INITIAL_ASSESSMENT.md
  4. ☐ Review 00_LINEAR_ALGEBRA_MASTER_PLAN.md
  5. ☐ Check 01_KNOWLEDGE_GRAPH.md for dependencies
  6. ☐ Set up NumPy or MATLAB
  7. ☐ Start Module 1.1!

🌟 Inspiration

"Linear algebra is the mathematics of data." — Gilbert Strang

"You can't do machine learning without linear algebra." — Every ML engineer

"The more I learn about linear algebra, the more I realize it's everywhere." — You, after completing this course!

Linear algebra is the foundation of modern technology. Master it and unlock AI, graphics, data science, and more! 📐🚀

Last Updated: October 21, 2025 Status: ✅ Complete learning plan Next Review: January 2026