Ex-NSA Hacker Is Building an AI To Find Hate and Far-Right Symbols on Twitter and Facebook (vice.com)
Motherboard reporter Lorenzo Franceschi-Bicchierai has interviewed Emily Crose, a former NSA hacker, who has built NEMESIS, an AI-powered program that can help spot symbols that have been co-opted by hate groups to signal to each other in plain sight. Crose, who has also moderated Reddit in the past, thought of building NEMESIS after the Charlottesville, Virginia incident last year. From the report: Crose's motivation is to expose white nationalists who use more or less obscure, mundane, or abstract symbols -- or so-called dog whistles -- in their posts, such as the Black Sun and certain Pepe the frog memes. Crose's goal is not only to expose people who use these symbols online but hopefully also push the social media companies to clamp down on hateful rhetoric online. "The real goal is to educate people," Crose told me in a phone call. "And a secondary goal: I'd really like to get the social media platforms to start thinking how they can enforce some decency on their own platforms, a certain level of decorum." [...]
At a glance, the way NEMESIS works is relatively simple. There's an "inference graph," which is a mathematical representation of trained images, classified as Nazi or white supremacist symbols. This inference graph trains the system with machine learning to identify the symbols in the wild, whether they are in pictures or videos. In a way, NEMESIS is dumb, according to Crose, because there are still humans involved, at least at the beginning. NEMESIS needs a human to curate the pictures of the symbols in the inference graph and make sure they are being used in a white supremacist context. For Crose, that's the key to the whole project -- she absolutely does not want NEMESIS to flag users who post Hindu swastikas, for example -- so NEMESIS needs to understand the context. "It takes thousands and thousands of images to get it to work just right," she said.
At a glance, the way NEMESIS works is relatively simple. There's an "inference graph," which is a mathematical representation of trained images, classified as Nazi or white supremacist symbols. This inference graph trains the system with machine learning to identify the symbols in the wild, whether they are in pictures or videos. In a way, NEMESIS is dumb, according to Crose, because there are still humans involved, at least at the beginning. NEMESIS needs a human to curate the pictures of the symbols in the inference graph and make sure they are being used in a white supremacist context. For Crose, that's the key to the whole project -- she absolutely does not want NEMESIS to flag users who post Hindu swastikas, for example -- so NEMESIS needs to understand the context. "It takes thousands and thousands of images to get it to work just right," she said.
Nothing can possibly go wrong with this. It has everything: NSA, hacking, white supremacists, reddit, AI. Definitely worth funding.
And the far left gets a pass!
How about a moderation of -1 pedantic.
What about far-left wing symbols of hate, violence, and oppression?
Antifa flags, socialist fist, hammer & sickle
After all, the Communists have killed far more innocents than the Nazis did.
It definitely needs to find a way to filter out ANYTHING that might be Xenophobic because that's always racist.
Unless it's about Russia.
Or unless it's about Boycott, Divest & Sanction which is the politically correct way to literally act like 1933-era Hitler but OK on the UC Berkeley campus so not a hate group.
Or unless it's misgynistic AND xenophobic remarks about Melania Trump because reasons.
Or unless it's antisemitism directed at Ivanka Trump because similar reasons.
Communism is both much deadlier and more socially-accepted than any other kind of hateful school of thought today.
Anything "fighting evil" that ignores images of Che Guevara and like symbols is simply partisan b.s.
In Soviet Washington the swamp drains you.
It's even better than that. Not only are they not drying to hide, they pop up whereever some offended snowflake deems them appropriate. Can't win an argument on merits? Call the other guy a Nazi and you're golden! I'll demonstrate:
1. American immigration laws should be enforced.
2. There is no evidence that police in America apply different standards to white citizens and non-white citizens.
3. Islamic fundamentalism motivates the violent act committed by many extremists and must be combated and defended against using within the framework of foreign diplomacy, foreign aid, military policy, and immigration policy.
4. Government benefits should only be provided to the demonstrably infirm or aged citizens and not be made available to able-bodied persons of working age.
5. Restrictions on the sale, ownership, or possession of firearms punish the law-abiding and do not make any dent in violent crime.
All of those are either factually true or present an opinion within the mainstream of acceptible American thought. How long will it take for someone to label one or all of them extreme and me an extremist beyond the pale of acceptable civil discourse.
I know..go figure, right?
I have actually been seeing of late, YouTube rants of people actually arguing that if you are of any non-white color they you by definition can NOT be a racist.
Seriously?
Geez....common sense has gone 101% out the door in the US.
Light travels faster than sound. This is why some people appear bright until you hear them speak.........
I have actually been seeing of late, YouTube rants of people actually arguing that if you are of any non-white color they you by definition can NOT be a racist.
Well, dredging strawmen from the bottom of the YouTube comment barrel is hardly epistemologically impressive.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
The famous statement "I disapprove of what you say, but I will defend to the death your right to say it" (incorrectly attributed to Voltaire) has always been the best summary of the principle and right of free speech. Given the sad state of civics education in the past three decades, we're seeing the result of this failure to learn this basic idea. When we put "feels" above principle, we get garbage like NEMESIS that ignores the far left crackpots and singles out the far right crackpots. People really need to listen to their elders' take on free speech and democracy and not be a sucker.
Comment removed based on user account deletion
Ugh, here we go.
https://www.vox.com/cards/poli...
https://www.vox.com/policy-and...
I hope these facts will persuade you to stop spewing lies.
"When information is power, privacy is freedom" - Jah-Wren Ryel
First of all I didn't use "trans" as a pejorative. I did use "crazy" as one though.
Since you call me a bigot/racist, exactly which race do you imagine me to be? That I've put you on the spot with that question should indicate that you're already wrong and that you've been caught. The only dog-whistles here are those in your own head that you've been conditioned by others to hear. In other words, you are the bigot - not me.
Ugh indeed.
1. Vox and Mother Jones (which one of your links cites) are partisan publications, not neutral reporters of fact.
That aside,
2. Disproportionate police use of force with and arrests of members of different population groups do not imply different enforcement standards unless there is parity in overall rates of crime among those groups. Black Americans commit more crimes per capita and live in higher-crime neighborhoods on average. Thus more encounters with police. And often for good reason. Your purported source of neutral facts tacitly assumes things like that don't matter and you tacitly assume that disparate outcomes implies disparate treatment. It doesn't. White people who charge at cops also tend to catch a bullet for their troubles.
3. Sloppy wording and lose definitions make for the best scatter plots, don't they? "Gun-related" deaths includes suicides. Whether you chose to count it in the same statistic as homicide by firearm is nothing other than that: a choice. Count it to pad the numbers one way. Exclude it to pad the numbers the other way. Half of "gun-related" deaths are suicides in the US and while I don't have numbers for the other countries on the plot, let's say the slope of that line drops by half.
The per-state chart is another cart-before-the-horse abuse of statistics. My statement was about gun restrictions. The chart is about gun ownership. The two are not the same, and while the correlation is weak to start with, the argument for causation is not there. Minnesota, Vermont, Maine, Iowa, Utah and Oregon have large populations and high gun ownership but lower gun deaths per capita than Colorado, Nevada, Maryland, and Florida which also have high populations but lower gun ownership. Then there's Chicago and Maryland and Camden NJ in particular which has strict local and state laws and insane homicide rates.
Actually there is abundant data confirming that minorities get different treatment by police than whites. For example epidemiological research says whites are slightly more likely to use weed than blacks, but law enforcement statistics show that blacks are roughly 3x more likely to be arrested for marijuana possession.
It is also true that police actually shoot roughly twice as many white people per year as blacks, but there are five times as many whites as blacks. This doesn't mean that every time a cop shoots a black man that race is a factor, but statistically it is bound to be a factor in a large number of shootings, although not in the simplistic way favored by many left-wing blogs on the topic -- although that probably happens at least some of the time. Assuming that the police are no better or no worse than society at large they must have enough racist sociopaths to produce at least a few shootings like that per year.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
Many people claiming to want equality actually don't want equality at all, they want inequality which favors them.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!