Our paper “Typical Error Exponents: A Dual Domain Derivation” has been just published on the journal IEEE Transactions on Information Theory
What is it about? Here is a sketch:
The reliable transmission of information is possible thanks to “channel codes”. For this reason, they are widely used in digital applications like mobile communications, server farms, #IoT, fiber-optic communications, deep space communications, Blue Rays and many more. Now, how easy is it to find a “good” channel code? It was known that, under some conditions, if you pick at random from all possible codes you are likely to get a good one. Ok, but how good? And what does “likely” mean? In our paper we make a step towards the answer to these questions. We quantify how good typical codes are, studying how quickly their error probability vanishes on any type of discrete communication channel.
The full version can be found here
(Work funded by: Beatriu de Pinós fellowship programme / Government of Catalonia, European Union’s Horizon 2020 research and innovation programme / Marie Skłodowska-Curie grant agreement No. 801370, European Research Council under ERC grant agreement 725411)