AI review queue

Review AI-suggested connections

AI proposes; humans decide. These connections were suggested by grounding on the graph — each shares a mental model across disciplines. Nothing here is published: accepting only moves a suggestion into review as “AI-suggested”.

AI suggestions

These are AI suggestions, not established knowledge. They exist to prompt a human reviewer, never to teach learners directly.

Generating grounded suggestions…

Knowledge review

The assignments that most warrant scientific review — flagged or disputed first, then analogies. Ordering follows explainable reasons, not a score.

Human scientific reviews performed: 0. This view prepares review; nothing is marked “reviewed” by code. Nothing here is externally validated.

Natural selection × Optimizationanalogyflagged for review
Shown first because: explicitly flagged for review; an analogy used across 17 disciplines — a correspondence, not an identity.
Traits that raise reproductive success are favoured, hill-climbing fitness over generations.
analogy
Where it stops: Evolution has no objective function and no optimizer — 'maximises fitness' is a descriptive analogy. Selection is myopic, can be trapped on local peaks, and the fitness landscape itself shifts with the environment.
Evidence & review
Mapping status: flagged for review.
- Evolution (4th ed.) — contextual · needs-check · strong evidence
Reinforcement learning × Variation & selectionanalogyflagged for review
Shown first because: explicitly flagged for review; an analogy used across 14 disciplines — a correspondence, not an identity; no source support yet.
Reinforces rewarded actions, selecting successful behaviours over time.
analogy
Where it stops: RL optimises expected reward for one agent within its lifetime; selection acts across generations with no designer. The shared idea is differential reinforcement, not identity.
Evidence & review
Mapping status: flagged for review.
No source linked yet — this is an editorial interpretation.
Epidemic spread × Diffusion & spreadanalogyflagged for review
Shown first because: explicitly flagged for review; an analogy used across 14 disciplines — a correspondence, not an identity.
Infection propagates through a population by contact, like a diffusing quantity.
analogy
Where it stops: Diffusion assumes local, gradient-driven spread; real epidemics also jump non-locally through travel and contact networks, which pure diffusion cannot capture.
Evidence & review
Mapping status: flagged for review.
- The Mathematics of Diffusion (2nd ed.) — contextual · needs-check · strong evidence
Gradient descent × Variation & selectionanalogyflagged for review
Shown first because: explicitly flagged for review; an analogy used across 14 disciplines — a correspondence, not an identity.
Selects parameter updates that reduce error — selection over model weights.
analogy
Where it stops: Gradient descent follows an explicit objective's gradient; biological selection has neither an objective nor a gradient. The correspondence is structural (variation + differential retention), not literal.
Evidence & review
Mapping status: flagged for review.
- Deep Learning — contextual · needs-check · moderate evidence
Supply and demand × Equilibriumanalogyauthored editorially (not independently reviewed)
Shown first because: an analogy used across 15 disciplines — a correspondence, not an identity.
Surplus pushes price down and shortage pushes it up until the two forces meet.
analogy
Where it stops: Market 'equilibrium' assumes many price-taking actors and clears only under those conditions; real markets can persist far from it (sticky prices, power, information gaps).
Evidence & review
Mapping status: authored editorially (not independently reviewed).
- Principles of Economics (8th ed.) — demonstrates this case · needs-check · moderate evidence