(nothing new but...)

What if no limit in number of trial, and if we want to know someone else's PIN?

if we try one-by-one, the order is N and if we exploit random search it's loG N
as shown in the Fig. 1
Our goal is to find an algorithm less order of complexity.


PIN 
  - one- by -one
  - rondome

FIG.

A-memory
   - wide plateau of 100 % fitness while elsewhere zero

Jeep

Hinton & Nowlan

discussion on reduced Even-n-parity

still now it is challanged and analyezed
for example mil et al (2005) ...
the author extend to two needles whose hight are different, i.e. local needle and 
global needle.




Even-n-parity

Fig.
fig of typical 2 cases of n=14 which 
... never find the solution



Q-computation

Fig.
plot number vs 2^n 
plot y=sqrt x

Q-walk
   intimidating 
   what are still open?


at the moment I am writing this paper ... NYT

---
Hinton and Nowlan wrote
"Baldwinism is best understood by considering an extreme (and unrealistic) case in which the
combinatorics are very clear." 
Later, however, this has been challanged for very sophisticated case [][][]