πŸ“ Math Tools Explorer

Post 3 companion Β· Vectors Β· Gradient Descent Β· ML Metrics

Material Feature Vectors

Enter two materials' feature vectors (up to 4 features each). See norm, dot product, and cosine similarity computed live.

Vector A (Material 1)
e.g. [a, Z_A, Z_B, Ξ”EN]
Vector B (Material 2)
e.g. [a, Z_A, Z_B, Ξ”EN]

Results

β€–Aβ€– β€” Norm of Material A
–
√(x₁² + xβ‚‚Β² + x₃² + xβ‚„Β²)
β€–Bβ€– β€” Norm of Material B
–
√(x₁² + xβ‚‚Β² + x₃² + xβ‚„Β²)
A Β· B β€” Dot Product
–
Ξ£ aα΅’ Γ— bα΅’
cos ΞΈ β€” Cosine Similarity
–
–

Live Gradient Descent on a Loss Surface

Watch gradient descent find the minimum of a simple loss function L(w) = (wβˆ’target)Β² + noise. This is exactly what happens during ML training.

Final w
–
Final Loss
–
Iterations
–
Try this: Set Ξ· = 0.9 β†’ watch the path overshoot and oscillate (like a bad SCF mixing parameter). Then set Ξ· = 0.05 β†’ very slow but stable convergence. Sweet spot: Ξ· β‰ˆ 0.1–0.3.

ML Metrics Calculator

Enter DFT values (true) and ML predictions β€” one per line. Computes MAE, RMSE, RΒ².

1. Two material feature vectors have cosine similarity = 0.99. What does this mean?
2. Your gradient descent loss oscillates up and down and never converges. What should you do?
3. Which metric penalises large errors more heavily than small ones?
SCORE
0 / 3