Critiques of mass surveillance and data-mining

Discuss

21 Responses to “Critiques of mass surveillance and data-mining”

  1. Alistair Muldal says:

    This is pretty much the laziest argument I’ve heard made against PRISM so far. I’m going to go ahead and assume that someone at the NSA knows how to construct a proper null statistical model…

    • Cowicide says:

      It doesn’t seem like you actually read the article.  It’s about corruption, not proper statistical models.

      Here’s the part you should have read before making your post haste ignorance:

      FTA:

      Daily Caller: So what are they doing with all of this information? If they can’t stop the Boston marathon bombing, what are they doing with it?

      Binney: Well again, they’re putting an extra burden on all of their analysts. It’s not something that’s going to help them; it’s something that’s burdensome. There are ways to do the analysis properly, but they don’t really want the solution because if they got it, they wouldn’t be able to keep demanding the money to solve it. I call it their business statement, “Keep the problems going so the money keeps flowing.” It’s all about contracts and money.

      Maybe next time RTFA?

      • Alistair Muldal says:

        Actually I did RTFA, all the way down to “… Likewise, Nassim Taleb writes:” (the bit that Corey quoted) – that’s the part I’m referring to.

        In fact, if you read Nassim Taleb’s original article in Wired (http://www.wired.com/opinion/2013/02/big-data-means-big-errors-people/), you’ll see it actually doesn’t even mention the NSA at all – it’s a (somewhat half-baked) critique of Big Data in general.

        • Cowicide says:

          Once again, try reading what Binney said (again) and critically think about the greater point.

          Once again, they don’t want to solve this problem; It’s about corruption.

          it’s a (somewhat half-baked) critique of Big Data in general.

          Speaking of half-baked, null statistical models (while useful) are overrated (typically by grad students).

          http://dynamicecology.wordpress.com/2013/05/13/null-and-neutral-models-are-overrated/

          Sifting through mountains of varied data like phone records, social interactions, email, gps data, search term history, web surfing history, etc. isn’t ever going to be easy and flooding it with random, bad data doesn’t help the situation.

          In other words, good old fashioned police work is more likely to ferret out people like the Boston bombers before they strike.  Mining millions of innocent people will not.

          And, once again, it’s not about security anyway, it’s about security theatre and corruption/money. You need to factor that in above all else.

          Also, I don’t enjoy living in a police state anyway where all our communications are monitored.  I’d rather risk some safety than live as a coward slave.

          • Alistair Muldal says:

            Seriously, did you just Google “null statistical model” and find the first hit with a negative-sounding title?

            Drawing reliable inferences from huge heterogeneous datasets is not easy, but it is absolutely wrong to think that it can’t be done. These issues are nothing new, and there are plenty of private companies that successfully apply machine learning algorithms to data on similar scales – think about Google’s “Search by Image” or Netflix’s uncanny ability to guess your taste in movies.

            Now consider that the NSA has billions of dollars and the best minds in the world to throw at the problem. *They* clearly think they can get something useful out of this data, and I would hesitate to bet against them.

            I’m more worried about the possibility that mass-surveilance *could* actually work.

          • Cowicide says:

            Seriously, did you just Google “null statistical model” and find the first hit with a negative-sounding title?

            No. And, did you bother to read it? No.

            So do you have a legitimate argument against the link I gave you or are you simply going to continue with ad hominem attacks?

            I’ve worked with business and government for various research projects and I can assure you that working on this vast amount of complex data isn’t going to remove false positives with null statistical models.

            Why don’t you show me where you could make it work? Put your money where your mouth is?

            there are plenty of private companies that successfully apply machine learning algorithms to data on similar scales

            Bullshit.

            [Citations sorely needed]

            And, please, it needs to be data that’s very similar to what the NSA is working with. You do understand that different data sets have their own separate challenges, don’t you? Researching the behavior of consumers who prefer extra large Big Macs is far different than combing through much larger mountains of data deciphering “terrorist-speak”.

            Drawing reliable inferences from huge heterogeneous datasets is not easy, but it is absolutely wrong to think that it can’t be done.

            I’m not saying it can’t be done, but you’re going to get a lot of false positives and it’s also wide open to corruption. That’s the point.

            I’m more worried about the possibility that this sort of mass-surveilance *could* actually work.

            Once again, it’s not meant to work. Hence, that’s why I keep referring you to William Binney (who actually worked within the NSA and with these type of programs).

          • Alistair Muldal says:

            No. And, did you bother to read it? No.

            So do you have a legitimate argument against the link I gave you or are you simply going to continue with ad hominem attacks?

            For the second time, yes I did read it.

            The thrust of this article is that because “the world is overdetermined”, null models are necessarily deficient because it is not possible to know precisely which underlying processes are ‘important’ with regards to the structure we see in our observed data. So null models are imperfect, but we’re talking about exploratory data analysis here, not hypothesis testing. The null model doesn’t have to be perfect, all that’s needed is a way to narrow down a large number of initial hits into a smaller number worth allocating greater resources to. Comparing the goodness of fit of a judiciously selected set of statistical models could go some way towards achieving this.

            [Citations sorely needed]

            You know as well as I do that such algorithms are going to be trade secrets, but to give Google as an example you can get some idea of what they are likely to use on the basis of the research they are funding:

            http://research.google.com/pubs/ArtificialIntelligenceandMachineLearning.html

            You do understand that different data sets have their own separate challenges, don’t you? Researching the behavior of consumers who prefer extra large Big Macs is far different than combing through much larger mountains of data deciphering “terrorist-speak”.

            Progress is incremental – the NSA don’t need to jump straight from Big Macs to “deciphering terrorist-speak”. Any way in which the database could be used to optimize the allocation of intelligence resources would be beneficial. For example, have you ever received direct communication with a terrorist or suspected terrorist, have you have ever contacted someone who has etc.

            And, please, it needs to be data that’s very similar to what the NSA is working with.

            What, like people’s phone records and the contents of their emails? What public entity could possibly have access to that kind of a database?

            Once again, it’s not meant to work. Hence, that’s why I keep referring you to William Binney (who actually worked within the NSA and with these type of programs).

            This assertion is what I really take issue with. It seems you’re only prepared to argue starting from the premise that the NSA cannot possibly be trying to do anything useful with the data it’s collected.

          • Cowicide says:

            The null model doesn’t have to be perfect

            Let’s see if you feel the same way when the false positive means you get investigated.

            Google as an example

            Not even close.

            What, like people’s phone records and the contents of their emails? What public entity could possibly have access to that kind of a database?

            Access isn’t the same as analysis.

            It seems you’re only prepared to argue starting from the premise that the NSA cannot possibly be trying to do anything useful with the data it’s collected.

            That’s untrue, but I’ll humor you anyway. It isn’t that they cannot, but they don’t want to in order to keep the funding pouring in. Sorry, but I’ll take William Binney’s word as a heroic whistleblower with direct experience at the NSA over your assertions any day.

            Also, I’ve made it clear that I think they can use the data, but it’s not worth all the false positives and living in a police state. Do you still not understand this concept?

        • Rindan says:

          I work with big data sets.  When you sift through big data, you got metric shittons of false positives.  This is okay.  While it makes “big data” not exactly the panacea that it has been touted as at times, that doesn’t eliminate its usefulness.  If you are looking for a needle in a haystack, tossing out a pile of not needles can be useful.  A little thought can eliminate a pile more of not needles.  You then have to take what you have left and find a way to test it.  You shouldn’t be shocked if in the end it turns out that you had nothing all along, but sometimes you strike it rich.
          Now, when I start sifting through my big data sets, the consequence of chasing down a false positive is some wasted time.  When the NSA gets a false positive it tears your life in half looking for signs of terrorism, or so we are going to infer, seeing as how their methods are secret.

          I don’t know about  you, but I REALLY don’t want a massive military spy agency sifting through all of my shit because I happened to correlate to some random profile.  Its a waste of all of our money, and it scares the living shit out of me because I don’t know what they are going to do with the information they gather.

          There are a lot of reasons to revile what the NSA has done in the name of stopping a threat that ranks below “murder by bathtub” (which is roughly 1 in 800,000 for anyone keeping store).  The fact that it is going to spit out false positive like a motherfucker is just one reason.

          The larger reason to be pissed is because spending billions of dollars and curb stomping liberty to fighting terrorism is stupid and cowardly.  Any American who is scared of terrorism, a threat that literally rates below death by bathtub, is a fucking coward.

          • theophrastvs says:

            (heh)well-done! “metric shittons of false positives”  (no units for false negatives? (“errors of type I”))

      • the happy drone says:

        Yup. It’s really all about the gravy train. Binney is right, and he should know…

  2. SedanChair says:

    LA LA LA I’M AN “ORDINARY AMERICAN”

    NONE OF THIS APPLIES TO ME

    LA LA LA

    • anwaya says:

      We’d just like you to come down to the station and answer a few questions about your reading habits and your favorite dessert.

  3. Humbabella says:

    One correction – this is not a dirty little secret.  If this is indeed a secret at all then it is giant, whopping lie.

  4. vonbobo says:

    I heard someone mention that if you need to find a needle in a haystack, you don’t need a larger haystack. Instead, you need a better needle finder.

  5. the happy drone says:

    For what it’s worth, here’s my critique (written in pseudo-brospeak, just in that kind of mood I guess): http://www.thehappydrone.com/2013/06/dear-nsa/

  6. JonS says:

    So:
    * pervasive surveillance is bad.
    * huge datasets are worse than useless (at least for the nominal task)

    Fine. How do we move on? How do we close Pandoras Box now they’ve not only opened the lid, but thrown away the key and smashed the fucking hinges?

    • Rindan says:

      Call people who are afraid of terrorist fucking cowards.  Once we realize that spending piles of money and liberty on terrorism is an act of extreme cowardice, we will stop doing it. That is how we move on.

      We need a cultural change that puts terrorism where it belongs in the hierarchy of “shit you should worry about”.  Bathtubs murder more Americans than terrorist each year.  If someone said that they could cut the number of bathtub deaths in  half by spending 100 billion dollars and tapping all the phone lines, most people punch that person in the face for even suggesting something so stupid.  Make the same proposition but replace the word “bathtub” with  “terrorist”, which kill fewer people than bathtubs, and suddenly it is a good idea.  People who think this way need to be punched, or at least relentlessly mocked for their cowardice.  Give them something to real worry about.

      Seriously, the answer is to call people out as being fucking cowards and articulate as clearly as you can your disdain for their rank cowardice.  I find it works shockingly well, especially on Republicans with a macho fetish.

  7. medontlivenoprahsworld says:

    The article segment about false positives reminded me of a problem with my bank in 2008.
    I was broke and out of work. I then found a job out of state. My checking account had routinely carried a low balance for about a year. Naturally, a big move like this needs a large amount of cash. I was lucky that my family could loan me the money until I was reimbursed by the company for the move.
    I hire movers and make the arrangements. I then move myself with a carload of belongings and await the action of the movers as they were scheduled to begin in another week. My family comes through and deposits the money into my account to cover the movers and my apt. deposits etc.
    I get a call from my family that a notice had arrived from my bank. I told them to open it and it was a notice that my account had been frozen for “unusual activity”. F**king really? I am moving out of state and the bank pulls this crap. I call them up, it’s being frozen for a week and they can’t switch it back. I was screwed. I had to scramble to borrow the cash all over again. Those ruddy bastards.
    This is what a false positive can do to your situation. It can really screw things over for you. Taken further, it could ruin your life or take your life.
    I have glossed over some details, but only in the interest of the focus of my story.
    -
    Thanks for letting me rant. I’d offer up some bandwidth tax, but all the trees are dead.

Leave a Reply