🤖 AI Summary
This work investigates the fundamental limits of reliable communication over finite-length deletion and insertion channels, aiming to establish tight upper bounds on the trade-off between blocklength and frame error probability. To address the computational intractability and looseness of existing general converse bounds—stemming from the absence of a suitable reference output distribution—the paper introduces, for the first time, an efficiently computable optimal reference output distribution tailored to arbitrary finite-input/finite-output channels. Building upon the general information-theoretic converse framework, the authors integrate probabilistic coupling techniques with refined blocklength analysis to derive a concise, analytically tractable finite-length converse bound. This bound strictly improves upon the classical binary erasure channel bound and yields significantly tighter capacity upper bounds under identical parameters, thereby providing a more accurate theoretical benchmark for code design over deletion and insertion channels.
📝 Abstract
We develop upper bounds on code size for independent and identically distributed deletion (insertion) channel for given code length and target frame error probability. The bounds are obtained as a variation of a general converse bound, which, though available for any channel, is inefficient and not easily computable without a good reference distribution over the output alphabet. We obtain a reference output distribution for a general finite-input finite-output channel and provide a simple formula for the converse bound on the capacity employing this distribution. We then evaluate the bound for the deletion channel with a finite block length and show that the resulting upper bound on the code side is tighter than that for a binary erasure channel, which is the only alternative converse bound for this finite-length setting. Also, we provide the similar results for the insertion channel.