The stated fast-contraction theorem and its input requirements
The manuscript asserts a fast contraction theorem for fixed constants C,A,b. With size parameter s>=2 and description length L, require: a bipartite network with at most Cs wires, maximum degree at most ceil(s^(9/10)), polynomially many vertices, and entries in B_(P0,ell) with 1<=P0,ell<=Cs. Each requested entry must be computable uniformly in time (s+L)^b 2^(A s^(9/10)) and space (s+L)^b. The stated contraction bound is (s+L)^{O(1)} 2^{O(s/sqrt(log s))} time and polynomial space. This is the theorem contract, NOT its proof. The recursive contraction proof remains outside this encoding-only learning-path stage; subsequent speed bounds are conditional on that theorem.
0
1
Tags
Math
Related
Subdividing wires makes a tensor network bipartite without changing its value
Why the sum of incidence-network contractions equals the lifted Boolean sum
Giving each equation-variable incidence its own polynomial coordinate
Representing the extended truncated coefficient ring at finite precision
The manuscript parameter recipe for a sparse measurement layout
Counting the auxiliary substitution and restriction pairs
Why operation counts need bounds on integer bit lengths
Bounding coefficients produced by the parity-lift construction
The quantitative contract for the initial sparse checks
Grouping tests and variables before indexing incidence subsets
Choosing enough residue-field elements for interpolation
Learn After
Related
Subdividing wires makes a tensor network bipartite without changing its value
Why the sum of incidence-network contractions equals the lifted Boolean sum
Giving each equation-variable incidence its own polynomial coordinate
Representing the extended truncated coefficient ring at finite precision
The manuscript parameter recipe for a sparse measurement layout
Counting the auxiliary substitution and restriction pairs
Why operation counts need bounds on integer bit lengths
Bounding coefficients produced by the parity-lift construction
The quantitative contract for the initial sparse checks
Grouping tests and variables before indexing incidence subsets
Choosing enough residue-field elements for interpolation