Candidate hypotheses · none proven · you decide

Engram Hypotheses

The idea hiding between two sources nobody read together, proposed by a machine, killed-or-kept by a jury, shown with its sources.

← latest

Frontier · open review

These are connections we are not sure about, shown on purpose. Some the system surfaced but could not ground: it searched the literature and still could not confirm a non-trivial, testable bridge. Others passed our jury and were published, until a closer, stricter look raised a real doubt. Either way they are open: not rejected, not confirmed.

Why show them at all? Because our jury is a set of models frozen at their training date, and a model cannot reliably validate something genuinely new, or always catch a subtle flaw. A person might know better. Read one, then make the case for it, against it, or link the source that decides it.

• Does a neuroevolution algorithm do best when it obeys biology's mutation-load law? ecology-evolution × machine learning
• Does a metal's slip-rate hardening law hide a temperature it never names? materials science × thermodynamics
• Is thermodynamic free energy and a solid's elastic strain energy more than a shared name? thermodynamics × materials science
• Are the brain's spikes and its graded signals two readings of one timing code? neural coding
• What if a learning agent settling on a policy is the same kind of process as a reaction reaching equilibrium? reinforcement learning × physical chemistry