In addition: Can decoding be performed by a device with light-computational power?
I understand that the answer might be a bit too long. In that case, I would appreciate some references with examples if possible. Thank you in advance.
This coding scheme really troubles me... It is the randomisation step for secrecy that confuses me. I will be specific:
See paper https://ieeexplore.ieee.org/abstract/document/6280261, page 3, column 2, paragraph 2 middle. The part of the paragraph looks like this:
"Alice randomly divides the typical h˜Δ 1,A sequences into non-overlapping bins, with each bin having 2mI(h˜Δ 1,A;h˜Δ 1,B) typical h˜Δ 1,A sequences. Hence, each sequence has two indices: bin number and index within the bin. Now, after observing the vector h˜1,A, Alice sets the key to be the index of this sequence within its bin. Alice then sends the bin number as the helper data to Bob through the public channel. That is, Alice needs to send H(h˜Δ 1,A|h˜Δ 1,B) bits of information through the public channel, where H(X|Y ) denotes the conditional entropy of X given Y . After combining the information observed from the public channel with h˜1,B, it can be shown that Bob can recover the value of h˜1,A with the probability arbitrarily close to 1. Then Bob can recover the key. "
My question is: how is possible for Bob to recover the key? Does he know the bin that Alice chose for her sequence?