Title:Identification Over Binary Noisy Permutation Channels

Abstract:We study message identification over the binary noisy permutation channel. For discrete memoryless channels (DMCs), the number of identifiable messages grows doubly exponentially, and the maximum second-order exponent is the Shannon capacity of the DMC. We consider a binary noisy permutation channel where the transmitted vector is first permuted by a permutation chosen uniformly at random, and then passed through a binary symmetric channel with crossover probability $p$. In an earlier work, it was shown that $2^{c_n n}$ messages can be identified over binary (noiseless) permutation channel if $c_n\rightarrow 0$. For the binary noisy permutation channel, we show that message sizes growing as $2^{\epsilon_n \sqrt{\frac{n}{\log n}}}$ are identifiable for any $\epsilon_n\rightarrow 0$. We also prove a strong converse result showing that for any sequence of identification codes with message size $2^{R_n \sqrt{n}\log n}$, where $R_n \rightarrow \infty$, the sum of Type-I and Type-II error probabilities approaches at least $1$ as $n\rightarrow \infty$. Our proof of the strong converse uses the idea of channel resolvability. The channel of interest turns out to be the ``binary weight-to-weight (BWW) channel'' which captures the effect on the Hamming weight of a vector, when the vector is passed through a binary symmetric channel. We propose a novel deterministic quantization scheme for quantization of a distribution over $\{0,1,\cdots, n\}$ by an $M$-type input distribution when the distortion is measured on the output distribution (over the BWW channel) in total variation distance. This plays a key role in the converse proof.