arXiv:2609.40211
Submodularity of entropy under quantum convolution
Milad M. Goodarzi
Abstract
We develop a submodular framework for the von Neumann entropy of discrete quantum convolutions, providing a noncommutative counterpart to the direct side of entropic additive combinatorics. We first introduce globally weighted quantum convolutions, which form compatible families indexed by admissible subsets of a fixed collection of inputs. Our main theorem reveals a polymatroidal geometry underlying their entropy growth: relative to any fixed admissible input block, the entropy gains admit a normalized, monotone, submodular extension to all subsets of the remaining inputs. The theorem yields convolutional strong subadditivity, quantum Ruzsa triangle inequality, and quantum entropic Plünnecke--Ruzsa inequalities for arbitrary input states. For repeated inputs, it gives sharp comparisons of entropy growth across admissible scales; in particular, the quantum doubling constant controls all higher admissible convolution entropies with optimal exponents. Together, these results bring submodular methods from additive combinatorics into the quantum setting and provide a systematic route to broad families of convolutional entropy inequalities.
AI Summary
The paper asks whether entropy growth under discrete quantum convolution obeys the submodular structure underlying classical entropic additive combinatorics. It treats arbitrary $n$-qudit states in odd prime dimension $d$, using globally fixed nonzero weights $q_i$. A subset $A$ defines a physical convolution $\mathcal C_A$ when $\sum_{i\in A}q_i^2$ is a nonzero square in $\mathbb F_d$. The main theorem shows that, relative to any nonempty admissible block $R$, the gains $S(\mathcal C_{R\cup J})-S(\mathcal C_R)$ extend to a normalized, monotone, submodular function on all remaining subsets, including inadmissible ones. Consequently, fractional covers bound collective entropy growth by weighted sums of smaller-block gains. Three-input consequences establish convolutional strong subadditivity and the balanced quantum Ruzsa triangle inequality for arbitrary states, resolving the stated conjectures under their coefficient conditions. For admissible equal-weight repetitions, $[S(\mathcal C_{[m]}(\rho))-S(\rho)]/(m-1)$ is nonincreasing across admissible scales. When balanced binary convolution exists, this gives $S(\mathcal C_{[m]}(\rho))-S(\rho)\leq(m-1)\log\delta_q[\rho]$, with optimal exponent even for diagonal one-qudit states. The proof converts convolution into Hadamard multiplication of positive Weyl characteristic kernels. Their normalized entropy equals $S(\rho)+\log(d^n)$. A single separable auxiliary state reproduces all admissible convolution entropies as marginal entropies up to this constant; strong subadditivity and nonnegative separable conditional entropy supply submodularity and monotonicity. The framework requires compatible global weights and arithmetic admissibility. Its repetition bounds concern normalized equal-weight convolutions, not general fixed-pair iterations. It does not automatically recover the classical difference-based Ruzsa triangle inequality or provide conditional, quotient, or inverse structural theorems.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.