Continuous FunctionStart learning
FoundationsChecking saved investigationReading browser-local route memory before showing a continuation.

Foundation Lab

Overparameterization & Generalization, Double Descent

GPT-4-class models are deep in the overparameterized regime: parameters ≫ training examples, yet generalize well

Concept 5 of 100OptimizationPhase 3
Open local demoAll Foundations
#5Double DescentOptimization
key equation
E[(y−y^)2]=bias2+variance+σ2\mathbb E[(y - \hat y)^2] = \text{bias}^2 + \text{variance} + \sigma^2
Reading map and next steps

Selected Foundation Object

Keep the equation fixed; move through the evidence.

Concept 5 of 100Double DescentOptimization / Phase 3: Optimization & generalization
Current question

Visual intuition for why larger nets can generalize better (not just "they memorize more")

E[(y−y^)2]=bias2+variance+σ2\mathbb E[(y - \hat y)^2] = \text{bias}^2 + \text{variance} + \sigma^2
PredictionCommit before the demo.

Ask what should change when the equation is manipulated, then let the local visualization test that expectation.

EvidenceCompare local witness and source.

Use the local runnable panel, the key equation, and canonical papers as separate witnesses for the same object.

InvariantName what survives notation changes.

The useful learning product is the reusable mechanism you can carry into another model, paper, or engineering tradeoff.

Next moveContinue in the newer notebook.

This atlas page keeps a local demo; the domain notebook carries the fuller Intuition -> Math -> Code -> Demo sequence.

Open local demoInspect equationOpen newer notebookNext foundation

Why It Matters for Modern Models

  • GPT-4-class models are deep in the overparameterized regime: parameters ≫ training examples, yet generalize well
  • Chinchilla-style scaling laws model how test loss scales with both capacity and data

What Tutorials Skip

What is still poorly explained in textbooks and papers:

  • Visual intuition for why larger nets can generalize better (not just "they memorize more")
  • How implicit biases of different optimizers (SGD vs Adam) select among infinite interpolating solutions

Interactive Visualization

Core Math (Optional Deep Dive)

If you want intuition first, start with the key equation and local visualization. Come back here for the full walkthrough.

Key Equation
E[(y−y^)2]=bias2+variance+σ2\mathbb E[(y - \hat y)^2] = \text{bias}^2 + \text{variance} + \sigma^2

Classical learning theory: test error ~ U-shaped function of model capacity. Empirically, modern nets show double descent: as parameters cross the interpolation threshold (0 training error), test error drops again as capacity keeps growing.

In simple linear models:

E[(y−y^)2]=bias2+variance+σ2\mathbb E[(y - \hat y)^2] = \text{bias}^2 + \text{variance} + \sigma^2

Variance explodes near interpolation, then decreases as overparameterization plus implicit regularization kicks in.

Canonical Papers

Understanding Deep Learning Requires Rethinking Generalization

Zhang et al.2017ICLR
Read paper →

Reconciling Modern Machine-Learning Practice and the Bias–Variance Trade-off

Belkin et al.2019PNAS
Read paper →

Connections

Prerequisites

⌇Sharpness

Enables

ΘNTK↗Scaling💡Grokking

Next Moves

Choose the next question to carry this object forward.

Semantic Connections

Break / Breaks When
↗Small data→Scaling∀Bias-variance puzzle→Theory⌇Beyond interpolation threshold→SharpnessΘRich regime dominates→NTK
Map / Analogy
⌇Geometry ↔ Generalization→Sharpness