Live data from Hacker News

The NSA’s machine learning algorithm may be killing thousands of innocent people

arstechnica.co.uk

131–140 of 221 posts

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#131
post #130

Earlier quoted context omitted.

Wait wait wait. So they're killing people based on a model that fails Statistics 101 !? No accounting for base rates? No cross-validation!? God fucking damnit, why the fuck does our government think killing should be cheaper and easier than healing!? We put new medical treatments through decades of difficult scientific tests, but have none at all for strictly invalid ways to kill people?

I do not think you completely understood the article.

I'm pretty sure I did. Random Decision Forest with no regularization, no probabilistic accounting for the base rate of the target class, and with the test and train sets mixed rather than performing proper cross-validation. They've almost definitely overfit, and almost definitely killed someone who wasn't actually a terrorist.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#132

They literally named it Skynet. They have an evil sense of humor. Actually using machine learning to detect terrorists isn't a terrible idea. But you are going to get an error rate, and probably a high one in the noisy real world. Maybe only 50% of the people you detect are actually terrorists. Maybe it's even worse than that. We can't even test it because there is no validation set and unreliable labels. The reasona…

Disingenuous title: NSA only collects data (SIGINT) and secures transmissions and identity of friendlies (SIGSEC). What others do with the data is not done by NSA.

Which is of course misdirection through diffusion of responsibility.

Stay classy.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#133

They literally named it Skynet. They have an evil sense of humor. Actually using machine learning to detect terrorists isn't a terrible idea. But you are going to get an error rate, and probably a high one in the noisy real world. Maybe only 50% of the people you detect are actually terrorists. Maybe it's even worse than that. We can't even test it because there is no validation set and unreliable labels. The reasona…

Reading discussions on drone strikes makes you wonder what ever happened to pacifism. This whole idea that you can just go around other countries killing people -- even if they intend on killing you, and even if you would be able to hit the right target 100% of the time-- it's just old. It's primitive. And that is not to blame the USA. At this point in history humanity, as a collective, has the maturity of an 11 year…

I don't see why pacifism is "more evolved." Violence is one of the most fundamental interactions between organisms, right up there with eating and sexual reproduction. Is eating "primitive?" Are you looking forward to a future where humans "evolve" out of needing to eat? Throughout history, prosperous and successful societies have been built on the back of military power (Egypt, Rome, Great Britain).

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#134
post #14

Earlier quoted context omitted.

My guess would be because Pakistan has had good relations with US since it's independence. "The United States today provides extensive economic, scientific, and military assistance to Pakistan... and is Pakistan's most generous donor of foreign aid"[1] [1] https://en.wikipedia.org/wiki/Pakistan%E2%80%93United_States...

> My guess would be because Pakistan has had good relations with US since it's independence. More often not. There was a period from the 1970s when Pakistan was embargoed because of its nuclear power + weapons programmes, mainly in cooperation with France. This pushed Pakistan into the Chinese camp for arms supplies where they have remained firmly lodged ever since. During the 1980s the USA did supply some silver-bul…

To add to that, DoD just approved sale of 8 F-16s to Pakistan (http://thediplomat.com/2016/02/us-approves-sale-of-8-f-16-bl...). The ball is in Congress' court to decide if this actually goes through... but ya. Pretty much the definition of realpolitik.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#135

They literally named it Skynet. They have an evil sense of humor. Actually using machine learning to detect terrorists isn't a terrible idea. But you are going to get an error rate, and probably a high one in the noisy real world. Maybe only 50% of the people you detect are actually terrorists. Maybe it's even worse than that. We can't even test it because there is no validation set and unreliable labels. The reasona…

did you post this on the corresponding reddit thread as well? if not then someone literally copy pasted your comment, which I have been noting a lot more lately on there.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#136
post #48

Earlier quoted context omitted.

The fact that 2,500 to 4,000 people have been killed in Pakistan according to a completely opaque process for classifying targets and blasting them from the sky--a process that operates entirely outside any law--ought to be considered absurd, shocking, frightening, and soul-crushingly inhuman. But it's not. We're arguing about with each other about bullshit. I'm incredibly appalled by that. That's why stories like th…

* who will then produce evidence to show that this particular program does exactly what it tries to to* No they won't, because they've been challenged to do such things repeatedly in the past and always failed. These people live in a foreign country and it's not like the US Govt dispatches a bunch of detectives and lawyers based on the results of this ML model. Get real. The intelligence is handed off to the CIA with…

> The intelligence is handed off to the CIA without revealing how it's generated, the CIA then says "we got a list of terrorist couriers from the NSA, let's go get em" and boom, off it goes.

> This is all incredibly well documented.

Then feel free to submit at least one source (or better yet, several) for this "well documented" fact that (a) the NSA does not due any due diligence into suspected terrorists after they get picked up by this meta data process; (b) the NSA does not indicate in any way to the CIA how the list was generated; (c) the CIA uses this information without performing any of its own due diligence on the targets; (d) the CIA then goes and kills everyone on the list without any further approval from DNI, SECDEF, POTUS, etc.

I'm no drone program apologist but let's get real here to suggest that there's no intelligence or due diligence into these operations is willfully disingenuous at best.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#137
post #104

Earlier quoted context omitted.

We need a military. All states do -- its part of their duty to their citizenry. The problem in the US is that our hippy-run academia are driving the smart people out of jobs in the military (and police force for that matter), and it's putting us in a terrible position. The only way to fix it is for smart people to start joining the military again.

I thought the problem is that, 25 years after its only serious opponent in the world disappeared, the US spends as much on the military as the next 7 countries - five of which are US allies - combined.

Luckily through expert (sarcasm) US negotiations and a 3d chess policy, Russia just declared a new cold war and has warned of WWIII over Sryia. Eastern European countries are begging for US soldiers to be stationed there while they ramp up their own military spending to dissuade Russian aggression. So it looks like all that military spending will be needed after all.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#138
post #13

Big data analysis + mass surveillance is a frightening prospect. Of course you can train software to look for 'terrorists', but you could also train it to look for: - whistleblowers - minority groups (e.g. gay people, particular religious beliefs, political affiliation) - political dissidents - journalists whose behaviour changes - personal vulnerabilities (affairs, mental health issues etc) Think what authoritarian…

Israel allegedly identifies homosexual Palestinians based on mass surveillance data and exploits them for it [1, 2]

[1] http://www.haaretz.com/opinion/.premium-1.617280

[2] http://www.vice.com/en_uk/read/gay-palestinians-are-being-bl...

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#139
post #129

So yes, a state would not be much of a state without some form of military. But why do we need separate states? Edit: this is a genuine question about the practicality of a concept (the "nation-state") that was invented during the Gutenberg era of the printing press, and how such a concept has become impractical in the age of instantaneous international interconnectivity and economic globalization. Downvoters: How ar…

You really want to know? Because of regression towards the mean. Having a single state would mean that the US couldn't afford to have a significantly higher standard of living compared to the rest of the world. In fact, 1st world countries would have to meet 3rd world countries somewhere in the middle. (Assuming you are living in the states) Would you be willing to live on a wage of $5000 a year? Edit: see here https…

You're still thinking in terms of 19th century paper-based processes when you attempt to estimate globally integrated wealth in terms of a nationally fragmented paper-based information system ($). Also, GDP is not indicative of net wealth, as it is a measurement which implicitly asserts the broken window fallacy.

Over the long term, breaking down these legacy barriers to trade and freedom of movement would necessarily uplift everyone economically by principle of synergy.

Synergy: the interaction of elements that when combined produce a total effect that is greater than the sum of the individual elements, contributions, etc.

>I'm going to give you an important kind of a picture. I hear a lot of people say "I don't like machinery and technology, it's making a lot of trouble." So we're going to take all the machinery away from all the countries of the world, all machinery, all the tracks and the wires, and the works and we're going to dump it all in the ocean. And you will discover that within 6 months, 2 billion people will die of starvation having gone through great pain. So we say "That's not a very good idea lets put all the machinery back where it was." Then, we're going to take all the politicians from all the countries around the world and we're going to send them on a trip around the sun, and you'll find we keep right on eating. And the political barriers now... scientists say very clear you could make the world work and take care of 100% of the people at a higher standard of living that anyone has ever known despite the increasing population, but you can't do it with the barriers, any more than you can try run a human organism with a wall between the ear, the eye, and the stomach. It is an organic whole, it is total industrialization.

-Buckminster Fuller

>Human history can be viewed as a slowly dawning awareness that we are members of a larger group. Initially our loyalties were to ourselves and our immediate family, next, to bands of wandering hunter-gatherers, then to tribes, small settlements, city-states, nations. We have broadened the circle of those we love. We have now organized what are modestly described as super-powers, which include groups of people from divergent ethnic and cultural backgrounds working in some sense together — surely a humanizing and character building experience. If we are to survive, our loyalties must be broadened further, to include the whole human community, the entire planet Earth. Many of those who run the nations will find this idea unpleasant. They will fear the loss of power. We will hear much about treason and disloyalty. Rich nation-states will have to share their wealth with poor ones. But the choice, as H. G. Wells once said in a different context, is clearly the universe or nothing.

-Carl Sagan, Cosmos

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#140

Earlier quoted context omitted.

Reading discussions on drone strikes makes you wonder what ever happened to pacifism. This whole idea that you can just go around other countries killing people -- even if they intend on killing you, and even if you would be able to hit the right target 100% of the time-- it's just old. It's primitive. And that is not to blame the USA. At this point in history humanity, as a collective, has the maturity of an 11 year…

I don't see why pacifism is "more evolved." Violence is one of the most fundamental interactions between organisms, right up there with eating and sexual reproduction. Is eating "primitive?" Are you looking forward to a future where humans "evolve" out of needing to eat? Throughout history, prosperous and successful societies have been built on the back of military power (Egypt, Rome, Great Britain).

We, as a human species, are supposed to evolve beyond our basic immediate needs.

We created societies, because we need to eat, reproduce, survive, without reverting to base violence. We created medicine, technology to improve our quality of life and reduce the chances that biological evolutionary pressure destroys us. The black plague came close to killing 60% of the European population. It is now eradicated.

What is the next step in our evolution as a species? Right now the biggest danger to the survival of humankind is... humankind itself. If we are to survive, not for a hundred years, but for a hundred thousand years into the future, we must learn how to cooperate without annihilating each other in the process.

Post reply on HN