Why distinct residue-field points permit interpolation over a ring
Choose d+1 distinct residue-field elements and lift them to binary coordinate vectors a_0,...,a_d in O. Their pairwise differences are units. Define L_i(T)=product_{j!=i}((T-a_j)/(a_i-a_j)). This polynomial is one at a_i and zero at every other a_j. For a polynomial P of degree at most d, P(T)=sum_i P(a_i)L_i(T). For uniqueness, a difference vanishing at all points is successively divisible by each monic T-a_i: unit point differences ensure each remaining factor still vanishes at the other points. Its degree is too small unless it is zero. This works over A_p despite zero divisors. Without unit differences it can fail: modulo 4, the nonzero linear polynomial 2T vanishes at both 0 and 2.
0
1
Tags
Math
Related
Fixing each tensor entry at its creation precision
Evaluating a local wire kernel without dividing by a nonunit
Computing an inverse by doubling its modular precision
A finite geometric expansion of each crossing kernel entry
Recovering the count from several short local expansions
Measuring only the exceptional swap directions
Related
Fixing each tensor entry at its creation precision
Evaluating a local wire kernel without dividing by a nonunit
Computing an inverse by doubling its modular precision
A finite geometric expansion of each crossing kernel entry
Recovering the count from several short local expansions
Measuring only the exceptional swap directions