cybics/crystal/inference.md

the tru reads the cybergraph and speaks back in numbers. this is inference — computing explicit knowledge from collective learning

the tri-kernel (diffusion, springs, heat) produces cyberank, karma, and syntropy — deterministic, verified in consensus, visible to all. structure emerges that no single neuron created: paths, clusters, hierarchies born from local links

the tru half of intelligence. its counterpart is learning. decentralized inference handles interdisciplinary data that no single agent can process — the cybergraph integrates knowledge across domains, and the tri-kernel extracts structure from the whole

see standard inference for the algorithm

discover all concepts

Homonyms

soft3/glia/docs/inference
The Cybergraph as Generative Model The cybergraph is the model. The tri-kernel focus distribution $\phi^*$ is the "weights" -- a stationary distribution over particles that encodes what the network collectively considers relevant. A query is a context vector that biases the focus field. The…
neural/trident/src/neural/inference
inference

Graph