Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Reproducing Hacker News writing style fingerprinting

antirez.com

Post reply on HN
Print view View original

41–50 of 168 posts

Re: Reproducing Hacker News writing style fingerprinting

#41

Post by andrewmcwatters »

alganet wrote: post #32
andrewmcwatters wrote: post #23

Well, well, well, cocktailpeanuts. :spiderman_pointing: I suspect, antirez, that you may have greater success removing some of the most common English words in order to find truly suspicious correlations in the data. cocktailpeanuts and I for example, mutually share some words like: because, people, you're, don't, they're, software, that, but, you, want Unfortunately, this is a forum where people will use words like…

That seems to be a misconception. The usage frequency of simple words is a powerful tell.

I can understand the nuance of your assertion, but looking at the data returned by these results suggests it's not really all that powerful at all.

There are so many people that write like me apparently, that simple language seems more like a way to mask yourself in a crowd.

Re: Reproducing Hacker News writing style fingerprinting

#42

Post by hammock »

tobr wrote: post #39
hammock wrote: post #20

The "analyze" feature works pretty well. My comments underindex on "this" - because I have drilled into my communication style never to use pronouns without clear one-word antecedents, meaning I use "this" less frequently that I would otherwise. They also underindex on "should" - a word I have drilled OUT of my communication style, since it is judgy and triggers a defensive reaction in others when used. (If required,…

> Again, my thought on good, interesting writing is that these are to be avoided. You mean, ”I think this should be avoided”? ;)

Nice one high five

Re: Reproducing Hacker News writing style fingerprinting

#43

Post by gostsamo »

Frieren wrote: post #33

It works for me. The accounts I used long time ago are there in high positions. I guess that my style is very distinctive. But I also have seen some accounts that seem to be from other non-native English speakers. They may even have a Latin language as their native one (I just read some of their comments, and, at minimum, some of them seem to also be from the EU). So, I guess, that it is also grouping people by their…

I discover 2 people in my top 20 who I can bet are from the same country as me and it is not a big country.

Re: Reproducing Hacker News writing style fingerprinting

#45

Post by Frieren »

shakna wrote: post #34
alganet wrote: post #22

Cool tool. It's a shame I don't have other accounts to test it. It's also a tool for wannabe impersonators to hoan their writing style mimic skills!

I don't have other accounts, but still matched at 85+% accuracy for a half dozen accounts. Seems I don't have very original thoughts or writing style.

My guess is that people from the same region and similar background will have more and closer "alters". So, if you are Californian-American then there is many people that will speak similar to you in HN. If you are a Satawalese speaker then you may be quite alone in your own group.

(The Satawalese language has 460 speakers, most of who live in Satawal Island in the Federated States of Micronesia.)

Re: Reproducing Hacker News writing style fingerprinting

#46

Post by alganet »

shakna wrote: post #34
alganet wrote: post #22

Cool tool. It's a shame I don't have other accounts to test it. It's also a tool for wannabe impersonators to hoan their writing style mimic skills!

I don't have other accounts, but still matched at 85+% accuracy for a half dozen accounts. Seems I don't have very original thoughts or writing style.

It's a fingerprinting tool, not a profiling tool. You can't draw such conclusions from it.

What a profiler would do to identify someone, I imagine, requires much more. Like the ability to recognize someone's tendency of playing the victim to leverage social advantage in awkward situations.

Re: Reproducing Hacker News writing style fingerprinting

#48

Post by alganet »

andrewmcwatters wrote: post #41
alganet wrote: post #32

Earlier quoted context omitted.

That seems to be a misconception. The usage frequency of simple words is a powerful tell.

I can understand the nuance of your assertion, but looking at the data returned by these results suggests it's not really all that powerful at all. There are so many people that write like me apparently, that simple language seems more like a way to mask yourself in a crowd.

You can definitely mask writing style. If you can do that only by using simple words, I am not so sure.

Re: Reproducing Hacker News writing style fingerprinting

#49

Post by albumen »

Joker_vD wrote: post #30
hammock wrote: post #20

The "analyze" feature works pretty well. My comments underindex on "this" - because I have drilled into my communication style never to use pronouns without clear one-word antecedents, meaning I use "this" less frequently that I would otherwise. They also underindex on "should" - a word I have drilled OUT of my communication style, since it is judgy and triggers a defensive reaction in others when used. (If required,…

> I prefer "ought to" I too like when others use it, since a very easy and pretty universal retort against "you ought to..." is "No, I don't owe you anything".

Are you saying there's a connection between "ought" and "owe"? All I see is "I don't want to hear any criticism".

Re: Reproducing Hacker News writing style fingerprinting

#50

Post by antirez »

layer8 wrote: post #37
antirez wrote: post #11

Earlier quoted context omitted.

Imagine the 2D space, it also has the same property! You have three points nearby, and a fourth a bit more distant. 4 best match is 1, but 1 best match is 2 and 3.

Good point, but the similarity score between mutual matches is still different, so it doesn’t seem to be a symmetric measure?

Your observation is really acute: the small difference is due to quantization. When we search for element A, that is int8 quantized by default, the code paths de-quantize it, then re-quantize it and searches. This produces a small loss of precision, like that:

redis-cli -3 VSIM hn_fingerprint ELE pg WITHSCORES | grep montrose

montrose 0.8640020787715912

redis-cli -3 VSIM hn_fingerprint ELE montrose WITHSCORES | grep pg

pg 0.8639097809791565

So why cosine similarity is commutative, the quantization steps lead to a small different result. But the difference is .000092 that is in practical terms not important. Redis can use non quantized vectors using the NOQUANT option in VADD, but this will make the vectors elements using 4 bytes per component: given that the recall difference is minimal, it is almost always not worth it.

Post reply on HN