Micro-LEDs boost random number generation | Hacker News Reader

Micro-LEDs boost random number generation

discovery.kaust.edu.sa

I am not surprised, as most things that generate light are generating photons using quantum effects, and thus are true random. Furthermore, CCD and CMOS detectors themselves have a quantum efficiency less than 100%, meaning they only detect a fraction of the incoming photons at random. So, a regular light bulb in front of a webcam is already a quite high-bandwidth source of true random numbers.

So, there is nothing revolutionary going on there, this paper is more about how to build a system with micro-LEDs and a photodetector and how to remove any inherent biases in that system, with the obvious benefit of being able to make something very compact.

That’s not actually what the paper is about and it does discuss that LEDs did random number generation isn’t new. It’s more about using GaN microLEDs to produce high volume real time random numbers at a very low component, cost, and power. The key emphasis in the paper is on the high rates of number creation - 9Gb/s.

Can someone explain to me how this is different than a simple noise generator based on a PN junction? As in, isn't this just amplifying noise and aren't there less sensational ways of doing nearly the same thing? Does measuring a photon with this method actually get you better randomness? I have some serious gaps in my understanding here and an ELI5 would be neat.

Measuring photons in this manner gives you the best randomness. It is effectively a quantum technique. A PN junction is (mostly) classical.

The specific mechanism is mentioned in the article:

https://en.wikipedia.org/wiki/Spontaneous_emission

> Although there is only one electronic transition from the excited state to ground state, there are many ways in which the electromagnetic field may go from the ground state to a one-photon state. That is, the electromagnetic field has infinitely more degrees of freedom, corresponding to the different directions in which the photon can be emitted. Equivalently, one might say that the phase space offered by the electromagnetic field is infinitely larger than that offered by the atom. This infinite degree of freedom for the emission of the photon results in the apparent irreversible decay, i.e., spontaneous emission.

The question is whether quantum mechanical noise could have a conceivable advantage over classical noise. I strongly suspect: no. Classical noise is already factually unpredictable, so the theoretical unpredictability (assuming no hidden variable theories I guess) of quantum noise doesn't add anything.

Classical noise is only unpredictable if you are lacking the necessary physical information to make an accurate prediction. Otherwise, it is always predictable.

Quantum noise is not based in any kind of physical information in the same way. It is intrinsically random. The "randomness" isn't merely a side effect of a bunch of physical phenomena. You cannot compromise a QRNG even if you had perfect knowledge of the state of every particle in the system over time.

https://www.jpmorgan.com/technology/technology-blog/certifie...

>Quantum noise is not based in any kind of physical information

This sounds like more of a limitation of the model you are using than a limitation of reality.

> Classical noise is only unpredictable if you are lacking the necessary physical information to make an accurate prediction. Otherwise, it is always predictable.

Since you are always lacking that necessary physical information, it is always unpredictable. If it were otherwise, we would already know whether hidden variable theories of quantum mechanics (which lack intrinsic randomness) are correct or incorrect. But we don't know that. So intrinsic randomness doesn't make a difference to us. So quantum noise is useless.

I've been told that reverse shot noise from a PN junction is quantum in nature.

It is possible for an electron to spontaneously gather enough voltage to break through a PN junction backwards. This shows up as a very noisy current measured in microamps.

--------

A forward bias PN junction might not be quantomly random. I'll have to research more. But a reverse bias PN junction is almost certainly quantum in nature.

---------

IMO, this is all just PN junction noise. Maybe LEDs are better than Zener diodes for noise. I'm pretty sure that noise characteristics are a guess and check methodology, it's all PN junctions of slightly different shapes after all.

I suspect that "better randomness" is not what this solves, but rather faster randomness.

A PN junction gives you only megabits/s of randomness at most.

This proposed method, if the article is correct, reaches gigabits/s.

But it could be because they are just using a large array.

I'm sure I'm overlooking something, but what's the real use-case for true random number generation at that fast of a rate? Even a few Kb/s of random numbers is enough to continually reseed a cryptographic pseudo-random number generator that will generate as much output as you want that's indistinguishable from true randomness. I suppose you aren't reliant on the security of the underlying cryptographic primitives then, but you're still reliant on the particular hardware RNG chip being implemented in a way that's free of bias even if the underlying physics principle is sound.

One thing is for sure, nature has no shortage of randomness. So indeed it seems difficult to find advantages of a new method.

In any case, a PRNG might be a no-go for many applications, out of principle.

And also, maybe a PRNG requires more power and die area than this new method?

ScholarlyArticle: "Micro-LED based quantum random number generators" (2025) https://opg.optica.org/oe/fulltext.cfm?uri=oe-33-11-22154&id...

Huh, I thought QRNG generally referred to quasirandom number generators but apparently the quantum use is a lot more common.

Still a bit unfortunate of a name clash since they're pretty much the opposite thing.

At some point quantum randomness may turn out to be quasirandom too. But until then, indeed the terminology is confusing. I suggest using "pseudo randomness" instead, PRNG.

Pseudo-random and quasi-random number generators are different though.

I guess the better term is just low-discrepancy sequence, which is also what Wikipedia uses.

My first question would be whether it's possible to influence the output via triggering power fluctuations on the motherboard - e.g. by running expensive code to cause the CPU/GPU to scale up.

Probably not. It's hard to guess, but they probably get a Poison Distribution https://en.wikipedia.org/wiki/Poisson_distribution in the detector, they may read only a few of the lower bits of the data, and then mix them in the entropy pool, with other sources. So the end result is quite unpredictable.

It's somehow similar to a random generator where you have 5 dices, roll them and then add to the entropy pool only if the total was even or odd. Changing the power is like forcing the system to use only 4 dices. It changes the probabilities a little, but not in a very controlable way, and with a good mixing in the entropy pool it's almost irrelevant.

I read the actual open access paper: https://opg.optica.org/oe/viewmedia.cfm?uri=oe-33-11-22154&s...

Note if you look at the paper, you notice a close but not entirely perfect normal distribution, but nothing you cannot fix with UDNs and Irwin-Hall. For reference how that is done you can read the bottom of this very useful RNG article: https://people.ece.cornell.edu/land/courses/ece4760/RP2040/C...

My overall verdict on the tech in OP is that it is amazingly promising!

Anybody have input on why this isn't a "Paper Tiger"?

Because it isn’t a legal threat.

Why would it be?

I usually stick to lava lamps

Lava lamps have been deprecated, Lava LEDs are the new standard

Fender amps here

Only useful for random numbers up to 11 though.

Only a few Fender amps have good noise with the 5C1 wide panel Champ being king and the 5F1 narrow panel Champ being a close second. Silvertones beat them out but there is too much noise for most.

Reply on news.ycombinator.com