[ddiq]

Racial profiling is just a slur for pattern recognition. If it turns out that 95% of Chinese scholars in the US are more loyal to China than to America, and that when given a chance, will illegally send information back to China, what are you supposed to do?

It's fine if you don't want to have a racial bias yourself, but it doesn't mean you need to ignore the fact that (in this hypothetical) the vast majority of Chinese are highly ethnocentric.

[nnq]

> Racial profiling is just a slur for pattern recognition

...?! Did you really write that?! Do you really believe that? WTF, man! Do you realize how that would sound in any other context? A few steps down this line and you'd use machine learning concepts to justify natzy ideas or stuff like that.

[ddiq]

Machine learning will find these patterns all on their own. They're not worried about being ostracized for seeing the world for how it is.

[phreack]

Machine learning is based on sample data, which time and time again has been proven to be biased, producing errors due to racism which then get fed back into the system and perpetuate the cycle. It is not infallible and it does not reveal truths that people are 'worried' about seeing, it's a reflection of current mindsets.

[alfromspace]

It seems like you're taking "reflective of a reality I don't want to acknowledge" to mean biased.

[chronic71819]

> It is not infallible and it does not reveal truths that people are 'worried' about seeing, it's a reflection of current mindsets.

And who are you to say if the machine's output (ie, mindset) is to be trusted?

Machines are much less prone to bias than humans, that's for sure.

[nnq]

...that's one of the reasons why I think 60% of the time we're applying machine learning "at the wrong end of the problem". We should:

(1) pick a set of goals we deem worthy

(2) use machine learning to find patterns of actions that push things towards our goals - that's why I think we should move as much as possible in reinforcement learning, and figure out ways to use RL for data-science too, both supervised learning and time-independent unsupervised learning are really dangerous and biasing tools, and I think they're sort of dead ends on the road to true-AGI

There's no such thing as "seeing the world for how it is". Looking for patterns in data using a tool/machine trained on data (that will be inherently biasedly selected) will just lead you to find the patterns that confirm your f up world view, out of the infinity of patterns that there always are in everything...

We need to figure out how to manipulate the world in order to shape it into the patterns that we'd enjoy more.

There's a sort of "Chinese perspective" in applied machine learning nowadays (which is probably not originating in China, but being associated with "ML empowered surveillance and social scoring" is easier labeled like this) that I think is very wrong and extremely dangerous... we should grow up out of it quickly, because later on properly purging it from our cultures will be a violent endeavor... Heck, I even hope we as species get a Mars base up as a backup if this conflict of world views goes hot because there's a high chance it will be very very hot, like in nuclear hot...

[alfromspace]

Pointing and sputtering is not an argument.

[onion2k]

If recruiting Chinese nationals stops working the Chinese state will just start recruiting other nationalities. Plenty of people would pop a USB drive in their work PC for $100k. The only reason they don't now[1] is that it's more expensive, there's a bigger risk of the person turning themselves in, and if caught it's more politically difficult.

[1] Assuming they don't already, which is a huge assumption.

[lordfoom]

>If recruiting Chinese nationals stops working the Chinese state will just start recruiting other nationalities.

Presumably this will be harder for them to do.

[traderjane]

If black Americans are more likely to commit local crimes and you are developing software for police to predict hotspots of crime, wouldn't race be another aspect of pattern recognition in this narrative?

And would it be wrong for banks or loan programs to optimize on such pattern recognition if it were found that ethnicity could improve their forecasting? Or for software companies to forecast employee effectiveness on similar pattern recognition?

[mrtksn]

It's not really a dilemma.

Do you know what will happen with such an optimization? The pattern will be reinforced and you will end up with something like a civil war. At some point, your company will suffer because people will be busy shooting each other.

The movement towards fixing discrimination did not emerge because old folks were bad and the new folks are better people, it emerged because people recognized the issues it creates. It adresses the same issues that anti-competitive practices of the monopolies create: destruction of the market for short term gains of a single company.

[JackFr]

Whether those things are wrong is a good question for the debating society.

What they are is illegal.

[ddiq]

Do you want your system to be "nice" or do you want it do be effective?

[JackFr]

When we use the coercive power of the state, it can only be with respect to the actions of an individual. Group punishments are completely unacceptable.

Quote Rev. Bayes till you're blue in the face, but profiling amounts to holding a group responsible for the actions of bad individuals.

[onion2k]

I want it to be effective, which is why racial profiling is a bad idea. Crime rates correlate to poverty rather than race. If you design your system to target race you'll get lots of false positives in middle class black neighbourhoods and you'll ignore crime in poor white neighbourhoods. Racial profiling doesn't work.

[ddiq]

This is just not true. The percentage of blacks in a population is a higher correlate for crime than income, poverty, or any other economic factors. This has been replicated many times.

edit: Here are some studies. I would have put them in a reply but I'm being rate limited. Any one is sufficient on its own, but all are interesting.

http://www.unz.com/article/race-and-crime-in-america/

http://2kpcwh2r7phz1nq4jj237m22.wpengine.netdna-cdn.com/wp-c...

https://link.springer.com/article/10.1007%2Fs10940-011-9134-...

http://www.sciencedirect.com/science/article/pii/S0160289611...

http://www.jstor.org/stable/591624?seq=1#page_scan_tab_conte...

http://www.sciencedirect.com/science/article/pii/S0160289609...

https://link.springer.com/article/10.1007/s12552-016-9164-y

[hyperbovine]

> Any one is sufficient on its own, but all are interesting.

You use technical terms like "sufficient" and "replicated", as though you actually understood how science worked, but then link to a pile of crap scholarship published in third-rate journals by cranks and racists. It's annoying. And dangerous. Worse, you appear to be completely unaware of the distinction between correlation and causation, and what policy implications that has. Kindly get a clue.

[ddiq]

Can you look at the data and come to a different conclusion rather than attack the paper ad hominem?

[piersadrian]

This opinion is heavily biased toward a white-nationalist narrative. For instance, the second linked “study” is actually a propaganda piece published by the New Century Foundation, an organization described by its founder, Jared Taylor, as “white-separatist”.

HN is not the place for this.

[ddiq]

Can you find something wrong with the research itself in any of the papers that I linked?

[FilterSweep]

I miss when HN throwaways were used to serve as whistleblowers.

Not for spreading "The Bell Curve" 90s research and unsourced white nationalist propaganda.

[linkregister]

Can you share one of those studies?

[hyperbovine]

So class profiling then ? This does not sound any better.

[dragonwriter]

> Applying your intelligence effectively is what it's all about.

They actually do correlate to race, beyond that which income accounts for, but you'd kind of expect that if law enforcement was racist against blacks and was perceived as such: policing is less effective in black communities both because of police attitudes toward blacks and vice versa. (Racial profiling—literally law enforcement targeting people based on race—obviously exacerbates this.)

[djsumdog]

During WW2, there was an internment camp called Manzanar, where the US put all the Japanese, even those with full American citizenship, simply because they were Japanese.

[Judgmentality]

This does nothing to address GP's point.

GP: "What if the data actually shows 95% of Chinese people will break US law when given the chance?"

You: "Did you know there used to be Japanese internment camps?"

I think this is an interesting article that discusses what "racism" means in society today:

https://slatestarcodex.com/2017/06/21/against-murderism/

[z2]

I think both GP and GGP are somewhat loaded questions. The first is a hypothetical question that, while possibly genuine, can have a scaremongering side-effect. The second (perhaps unnecessarily) takes this as flamebait and extrapolates that cold calculation can lead to seemingly rational decisions such as mass interment based on the perpetual foreigner status of East Asians.

I have not studied rationalism or humanism at all and don't have much opinion on anti-discrimination or anti-anti-discrimination. The issue that sometimes boils over on mostly homogeneous countries like China is that people conflate many different things together. A distinction needs to be made, rather than blurred, on ethnic background, nationality, current government, their culture, or something else. (This is an assumption.)

Say what you will about 'The Chinese,' but please make it clear who and what you are talking about. A communist party member is very different from a middle class worker who holds a Chinese passport, or someone who has resided outside of China for decades. or a second generation immigrant with a Chinese-sounding last name. The above people probably have very different views on ethnocentrism or loyalty to 'the motherland' and should never be lumped together by proxy and acted on with uniform policy. This isn't racism because there are many other things at play, and discriminatory outcomes might defined by consequence. But such levels of ambiguous smearing can cause the same kind of prejudice from applying abstract, ill-placed fears more broadly than deserved.

[ddiq]

It's a perfectly reasonable stance in a war. Better to have 10,000 people interned than to risk a few more Takeo Yoshikawas running around.

[z2]

"While the Yoshikawa case appeared to retroactively justify the decision to intern Japanese Americans, he himself distrusted the Japanese-American community which in his mind was loyal to America over Japan."

https://en.wikipedia.org/wiki/Takeo_Yoshikawa#cite_note-Time...

[blub]

Yoshikawa was an embassy employee and had a military background...