Thinking·OS

Graph explorer

Explore the knowledge graph

Bias-variance tradeoff explains Model evaluation. Activate to inspect this relation.Classification is a Supervised learning. Activate to inspect this relation.Clustering is a Unsupervised learning. Activate to inspect this relation.Cross-validation is a Model evaluation. Activate to inspect this relation.Cross-validation measures Overfitting. Activate to inspect this relation.Dimensionality reduction constrains Overfitting. Activate to inspect this relation.Dimensionality reduction is a Unsupervised learning. Activate to inspect this relation.Feature engineering enables Supervised learning. Activate to inspect this relation.Gradient descent requires Loss function. Activate to inspect this relation.Gradient descent applies to Optimization. Activate to inspect this relation.Gradient descent depends on Training data. Activate to inspect this relation.Model evaluation depends on Loss function. Activate to inspect this relation.Optimization landscape applies to Gradient descent. Activate to inspect this relation.Overfitting depends on Training data. Activate to inspect this relation.Overfitting is a Trade-off. Activate to inspect this relation.Overfitting is part of Bias-variance tradeoff. Activate to inspect this relation.Overfitting suppresses Generalization. Activate to inspect this relation.Regression depends on Gradient descent. Activate to inspect this relation.Regression is a Supervised learning. Activate to inspect this relation.Regularization constrains Overfitting. Activate to inspect this relation.Regularization is part of Loss function. Activate to inspect this relation.Supervised learning depends on Loss function. Activate to inspect this relation.Supervised learning depends on Training data. Activate to inspect this relation.Supervised learning enables Generalization. Activate to inspect this relation.Supervised learning enables Prediction. Activate to inspect this relation.Supervised learning is part of Machine learning. Activate to inspect this relation.Unsupervised learning is a Machine learning. Activate to inspect this relation.Unsupervised learning is part of Machine learning. Activate to inspect this relation.Gradient descentOptimizationLoss functionTraining dataOptimization landscapeRegressionModel evaluationRegularizationSupervised learningOverfittingBias-variance tradeoffCross-validationMachine learningGeneralizationPredictionClassificationFeature engineeringTrade-offDimensionality reductionUnsupervised learningClustering
Relationship types

17 concepts viewed through this lens. Bridge concepts connect this view to Artificial Intelligence, Biology, Biostatistics, Business….

Legend
  • Focused concept
  • Connected concept
  • Bridge concept (just outside the lens)
  • Arrow points from cause / source to effect / target
  • A line with no arrow is a two-way relationship
  • Node colour marks the concept’s primary discipline
21 concepts28 relationships19 disciplines5 relation families

Gradient descent

Open concept →

At a glance

Gradient descent is a method for unconstrained mathematical optimization.

Disciplines
Computer Science · Machine Learning · Optimization · Data Science
Role in the graph
Cross-disciplinary bridge — reaches Business, Economics, Engineering, Mathematical Modelling, Statistics, Systems Engineering
Relationships
5 · 2 relation families
Mental models
3

What am I looking at?

In this lens (17)

Bridge concepts (21)

Just outside the lens — they connect it to other context.

  • Training data3
    Computer Science, Machine Learning, Statistics · connects to Gradient descent, Overfitting, Supervised learning
  • Generalization2
    Cognitive Science, Machine Learning, Statistics · connects to Overfitting, Supervised learning
  • Machine learning2
    Artificial Intelligence, Computer Science, Machine Learning, Statistics · connects to Supervised learning, Unsupervised learning
  • Binary numeral system1
    Computer Science · connects to Data
  • Causation1
    Epistemology, History, Medicine, Metaphysics, Philosophy of Science, Statistics · connects to Correlation

Insights from this view

Structural observations about the concepts shown here — descriptions of this graph, not claims about the world.

  • This view connects 19 disciplines: Artificial Intelligence, Biology, Business, Cognitive Science, Computer Science, Data Science, Design, Economics, Engineering, Ethics, Machine Learning, Mathematical Modelling, Mathematics, Optimization, Philosophy of Science, Political Science, Public Policy, Statistics, Systems Engineering.
  • Gradient descent is a bridge concept — viewed here through Computer Science, Data Science, Machine Learning, Optimization.
  • The connections here span 5 relation families.
  • Trade offs explains 3 concepts in this view (Biology, Business, Data Science, Design, Economics, Engineering, Ethics, Machine Learning, Mathematical Modelling, Optimization, Political Science, Public Policy, Statistics, Systems Engineering).

Relationships as a list

The focused concept’s relationships. Pick another concept in the graph above to update this list.

Explore through a different lens

A lens is a deterministic projection of the graph. Pick a discipline, thinking pattern or journey to reframe the whole view.

By discipline

ChemistryBiologyPhysicsMathematicsComputer ScienceGeologyMineralogyEconomics

By thinking pattern

Bayesian updatingCause and effectCompoundingConservationConstraintsDefense in depthDiffusion & spreadEmergenceEnergy transformationEntropy & irreversibilityEquilibriumFeedback loopIncentivesInformationLevels & hierarchyModularityNetworksOptimizationOscillationProbabilityRecursion & self-similarityScaleSignal vs noiseSource–pathway–receptorStocks and flowsSymmetry & invarianceThresholds & tipping pointsTrade-offsTriangulationVariation & selectionWaves

By journey

From a Leaf to the ClimateWhy Music Moves UsHow Machines LearnReasoning about a Leaf and the Climate

Concept collections

Concept collections are curated lenses onto the fabric — themed sets of ideas that recur across disciplines. They are not journeys; they are a way to read the graph.

About this view

What this is

Start from one concept and expand outward. The view never shows everything at once — click a node to refocus, filter by relationship type, or switch to an accessible list.

One fabric

3750 concepts and 5051 typed relations form one connected component — no isolated silo.

How to read it

Focus a concept, or apply a lens (discipline, mental model, journey) to see only the threads that matter.