finding

active

finding:learned-game-of-life-circuit-uses-336-active-gates-excluding-pass-through-gates-a-and-b-predominantly-or-and-and

Learned Game of Life circuit uses 336 active gates (excluding pass-through gates A and B), predominantly OR and AND

Circuit analysis result for Game of Life experiment

Source paper

extracted_from

Differentiable Logic Cellular Automata: From Game of Life to pattern generation with learned recurrent circuits

Neighborhood — ranked by edge-count

Claims (1)

claim

Different tasks yield distinct distributions of logic gate types between perception kernels and update networks
supports
Interpretive claim based on circuit analysis across experiments

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Learned Google G circuit uses 927 active gates (excluding pass-through gates A and B)finding0.853
Circuit complexity metric for colored pattern experiment, largest of all experiments
Learned lizard pattern circuit uses 577 active gates (excluding pass-through gates A and B)finding0.824
Circuit complexity metric for lizard experiment
Learned checkerboard generation circuit reduces to just 5 active logic gates after pruning (6 with one redundant AND)finding0.821
Remarkably minimal circuit discovered for checkerboard pattern generation
Total active gates in checkerboard circuit is 22 (excluding pass-through gates A and B)finding0.816
Circuit size finding for checkerboard experiment before pruning
Game of Life learned circuit successfully replicates gliders, stable blocks, loaves, and boats on larger grids beyond training sizefinding0.784
Demonstration of generalization of learned GoL rules to larger grids
Active Inference Rule-Learning Simulation (32 trials, 64 agents)method0.712
Computational simulation method using spm_MDP_VB_X to demonstrate curiosity and insight
DiffLogic CA fully converges on Game of Life rules — both soft and hard losses converge to zero on all 512 configurationsfinding0.701
Core result of Experiment 1 validating DiffLogic CA's ability to learn discrete CA rules
Circuits could act as an epistemic foundation for interpretability by breaking down model behavior into falsifiable statements about small subgraphs.claim0.698
Normative vision for how the circuits agenda could resolve the pre-paradigmatic state of interpretability