Disease Classification

In[]:=
rrules={4201857261303,4201857263490,4201857263409,4201856731968};
In[]:=
Table[SeedRandom[2424+i];PlotDifferences2[CellularAutomaton[{#,3,1},{{1},0},{200,All}],PerturbedCAEvolution[{#,3,1},{{1},0},200,40->1],"Trim"->{2,None}],{i,30}]&/@{4201857261303,4201857263490,4201857263409,4201856731968}
Out[]=

,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,

In[]:=
With[{ru=4201857263409},Table[SeedRandom[2424+i];PlotDifferences2[CellularAutomaton[{ru,3,1},{{1},0},{70,All}],PerturbedCAEvolution[{ru,3,1},{{1},0},70,40->1],"Trim"->{2,None}],{i,30}]]
Out[]=

,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,

In[]:=
With[{ru=4201857263409},Table[SeedRandom[2424+i];HighlightPerturbations[Reap[PerturbedCAEvolution[{ru,3,1},{{1},0},70,40->1]][[2,1,1]]],{i,30}]]
Out[]=

,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,


When it’s stretched to the limit, using all rule elements, it’s more likely fail completely when something is changed

Is that a measure of redundancy? [I.e. can you use different cell values and get “the same results”]

When do you see a region of change, with the same pattern all around it?

Once the change is healed across a whole row ... it will stay healed until there’s another perturbation [ in that case, it’s “masked” the perturbation ]

Therapies are like Maxwell’s demons (trying to put Humpty Dumpty together again...)

If the system is modular, maybe it’s like the Titanic, with “firewalls” in between parts

Does a periodic change in the fitness function lead to modularity?
(Long computation time + 8 GB)