High School Rare Entropy and Hidden Markov

[Kasimir]

when selecting rare entropy sources for trng and one can see similarities trough an applied hidden markov model, will it be still good entropy?
(structure is the same, even though type of source input is different)

 

[mathman]

Fix typos! trng??

 

[WWGD]

Fix typos! trng??

I know that rng is usually" Random Number Generator" but I am not sure about the t part.

 

[Klystron]

"True" random number generator: physical source of unpredictable timing pulses. Example, scintillation counter outputs generate the next seed value for a programmable or numeric prng.

Guessing "trough" corrects to "through". Also assuming "similarities" refer to comparing output states of the various "true" or hardware random number generator sources via hidden Markov rules?

 

[MarkOW]

I think he does not need rules. It seems a more basic question about precoditions.

 

[Klystron]

I think he does not need rules. It seems a more basic question about precoditions.

Thank you.

I know that rng is usually" Random Number Generator" but I am not sure about the t part.

Found a low probability match for string 'rng' from algebra meaning a ring without identity matrix, not a typical beginners topic?
Op needs initial (algebraic) seed conditions?From IT using 'information entropy' sum (p(log(p)) the seed can be selected from the set of expected output values. Crypto requires an input vector seed. Programmers often use synchronization (clock) pulses as (non-crypto classical) seed sources, riding the error rate (so to speak).

 

[MarkOW]

"Programmers often use synchronization (clock) pulses as (non-crypto classical) seed sources"

that's exactly the point. now, we do got two such input sources and apply hidden-markov. Is entropy better?

 

[Klystron]

Please define 'entropy' as used in question,"better than" what?.

Two homework sources can be combined and compared. Perhaps measure difference of the values, the interval between pulses or variation in pulse arrival from the expected value, to seed your prng. Design depends on requirements. As expressed on your sw/os/hw platform.

I understand commercial functions combine or mix rng outputs and error-check each value in the output buffer, filtering predictable strings. Found good info searching for <random seed> <information theory: entropy> and other concatenations.

 

[Klystron]

Von Neumann error correction provides measures while potentially increasing 'randomness'.
https://en.wikipedia.org/wiki/Von_Neumann_entropy

Error correction can operate above bit level described in the text but bit flip functions are fast and intrinsic. [citation?]
Emulating a pseudo-Markov-chain to test output buffer contents not difficult particularly if you already use a forward_link (flink) and backward_link (blink) and other pointers. Recursive rng functions workable at least at low frame rate, low output rates, slow applications; also limited by stack but solves 'seed source'. Recursion indicates function_trng () self-referential, also well behaved.

 

[Kasimir]

"Programmers often use synchronization (clock) pulses as (non-crypto classical) seed sources" that's exactly the point. now, we do got two such input sources and apply hidden-markov. Is entropy better? "

I meant cryptographic seed sources and trng. It seems MarkOW was irritated.