Indeed, and success/failure may give us an indication of how much more inherently contagious Omicron actually is, versus how much of our spread has been due to vaccine evasion and restriction relaxation.
Kolmogorov complexity/entropy is more suitable for this purpose, under the implicit assumption that password crackers don't have tailored prior knowledge and are just enumerating "simple" sequences. It only agrees with Shannon entropy on long ergodic sequences. The author basically constructed an example where the two notions don't agree.
Kolmogorov complexity is only unambiguously defined asymptotically, and "asymptotics is merely a heuristic". It is also uncomputable. So, to use entropy arguments for passwords, the only correct way I could think of is to generate long and (elementwise) random passwords.
You asserted that Kolmogorov complexity will disagree with Shannon entropy in this example, so how do you know what the Kolmogorov complexity of this example is?
It is a random variable in this setting, as it is a function of the randomly generated password. Given a deterministic sequence, you find the definition of its Kolmogorov complexity in textbooks/Wikipedia/etc. By saying the Kolmogorov complexity will disagree with Shannon entropy, I meant the former, which is a random variable here, does not converge to the latter, contrary to the standard asymptotic setting which probably gives people the idea of using entropy to characterize password (I don't know, don't work in security).
The point of my original post is that the asymptotics break down here, and this phenomenon is not poorly understood, at least in some other communities. It is not meant to provide an alternative that is always well-defined and useful, although as I said in the grandparent comment, there is the useful implication that you can stay safe by sticking to the asymptotic regime.
It's easy to calculate an upper bound on the Kolmogorov complexity of a given value (you just have to exhibit a Turing machine that computes it), but very hard to prove a lower bound (you would have to prove something about all possible Turing machines).
By giving the password to a good compressor? (and then computing the shannon entropy of the result) Yet i'm not sure i know a good compressor for short strings... Perhaps something like gpt2tc tailored to passwords instead of english text.
The implicit assumption however isn't good. Password crackers regularly make use of prior knowledge. A password that consists of a Shakespearian Sonnet for example has very high complexity but makes for a bad password.
Kolmogorov complexity kinda does account for "prior knowledge" (that's why it's not computable). A shakespearian sonnet will have low kolmogorov complexity (there's redundancy).
> On analysis of the file, I found that the vast majority of the records are actually related to sex, porn, and other smartphone brands. There are mentions of Tibet, Hong Kong, and other religious groups, however, mentions of the CCP and “China” are also included, too
> I think it’s pretty clear that the filter is specifically used for filtering advertisements
The article in the OP covers this directly. A thousand entries with normal advertisement-filter-sounding terms were added, and then everything removed.
I feel compelled to point out that you would know this had you read the article.
This is false. The ad keyword blocklist in the original report had tons of unassuming keywords. The authors cherry-picked a few to support the incredibly boring "China bad" narrative that was probably their goal.
That fell flat, so now there's a new rumor that the See See Pee got caught red-handed and disguised their Totally Real Evil Scary Censoring Algorithm Which Doesn't Work with unassuming keywords.
> Would censor approximately 450 words and phrases, it said. The blocklist wasn’t active, but could be activated remotely. It was filled with political terms, including “Democratic Movement” and “Long live Taiwan’s independence.”
> After the government published its findings, things got weird. The list swelled to more than 1,000 terms, including hundreds of non-political terms like “pornography,” seemingly to turn the political blocklist into something more generic. Then, it disappeared. “They reacted,” Margiris Abukevicius, Lithuania’s vice minister for defense, told me. “It wasn’t publicized from their side.”
So it seems you're the one pulling some mental gymnastics.
> This is false. The ad keyword blocklist in the original report had tons of unassuming keywords.
There is no statistical good/bad measurement if you, as a company, decide to block terminology of political nature.
They blocked pro-Taiwan ads just as they blocked outdated Xiaomi ads. But this doesn't even out the odds anyhow. As long as pro Taiwan ads are blocked, it's effectively being used as a censorship mechanism, and it's nothing else.
Doesn't matter if you sell it as an adblocker. Doesn't matter if you sell it as a filtering system for "security" reasons.
Either block all ads and do the right thing, or block no ads at all. But not this.
The point is that they realized that advertisements are used to organize democratic or revolutionary movements and they blocked it. And that is the censorship others are complaining about.
So what that it has unassuming keywords? If they thought the list might get out, then of course they would add a lot of keywords that don't matter. The computer doesn't get tired because it now has to check for 100x the number of keywords.
> Steiner is not prepared to rest on his laurels. He is currently reworking part of his dissertation for publication and plans to continue his theoretical physics work.
It definitely helps that he has a lifetime of research experience (in another field).
I was about to post the same thing. Self-organisation maps seem a classical computational model to me. If the author’s point was that computational models should be biologically plausible, there are many other examples as well. I’ve never really understood what neuroscientists are talking about…
The nice thing about the amortisation strategy is that you can e.g. exercise on the slow days, which improves your stamina and thus efficiency in the long run.
