concept
active
concept:asynchronous-updatesAsynchronous Updates
Update strategy where random subsets of cells are updated per step, simulating independent cell clocks
Neighborhood — ranked by edge-count
Methods (1)
method
- Asynchronous Update TrainingimplementsTraining regime where random subsets of cells update per step, improving robustness of learned circuits
Concepts (2)
concept
- Synchronous Updatescontradictsrelated_toDefault update strategy where all cells update simultaneously; contrasted with asynchronous updates
- Fault Tolerance and Self-Healingassociated_withDiffLogic CA circuits exhibited emergent fault tolerance without explicit design for it
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Communication where sender does not block; dominant in Linda.
- Second stage of DiffLogic CA where a DLGN computes each cell's new binary state from perception output and current state
- Models trained directly with asynchronous updates would exhibit even greater robustness than synchronously trained modelshypothesis0.721Hypothesis that motivated the asynchronous robustness comparison experiment
- Blocking remote procedure call; Linda can implement it but it is less natural.
- TUAI failure mode: world model basis shifts relative to prediction basis, rendering predictions nonsensical noise.
- The capability of an evolver model to produce useful persistent harness updates from execution evidence
- Basis for preferring Linda's out (asynchronous) to remote procedure calls.
- Standard method in mechanistic interpretability that intervenes on activations; VPD flips this paradigm by patching parameters.