Anonymity of Netflix Prize Dataset Broken

KentuckyFC writes "The anonymity of the Netflix Prize dataset has been broken by a pair of computer scientists from the University of Texas, according to a report from the physics arXivblog. It turns out that an individual's set of ratings and the dates on which they were made are pretty unique, particularly if the ratings involve films outside the most popular 100 movies. So it's straightforward to find a match by comparing the anonymized data against publicly available ratings on the Internet Movie Database (IMDb) (abstract on the physics arxiv). The researchers used this method to find how individuals on the IMDb privately rated films on Netflix, in the process possibly working out their political affiliation, sexual preferences and a number of other personal details"

Comments

Sexual preferences?

[tygerstripes]

Who goes out of their way to rate "Anal Whores 3" online?

Re:Sexual preferences?

[morgan_greywolf]

Bill Clinton?

Re:Sexual preferences?

[mh1997]

Who goes out of their way to rate "Anal Whores 3" online?

The good thing about porn flicks, as a general rule, is that they're too bland to have really bad plots. The search for good dialogue strays too far off the beaten path established by the social mores of the target market, be that old men, college students, or perverts out on dates. There are pornos with solid plots, just rarely pornos with complicated plots.

What they generally aren't is full of capers designed by crackheads in search of sexual relief, or a dominatrix dying to destroy the gold market with a Da Vinci alchemy machine only a cat burglar from Hoboken could steal.

Yes, the plot of Anal Whores 3 is as convoluted as it is kitschy. Mercedes and Veronica Diamond forcibly enlist the help of happy-go-lucky and half-a-second-out-of-prison pizza delivery man Hawk (Peter North) to steal the pieces to a machine that turns lead vibrators into gold. Hawk isn't halfway to a cup of coffee with his wise cracking cohort, Tommy (Johnny Cockring) when he finds himself back in the burglary game. Casing out a heist he meets nun/professional patron of the arts/double agent/love interest Jessie Jane (vows of bestiality can put the kibosh on even the best of cinematic love interests). When you throw in a CIA agent (Dick Coburn) and a couple of double dildos, you've managed to make the world's most convoluted porno....

Re:Sexual preferences?

[styryx]

That's the plot of Hudson Hawk. Good flick.

Re:Sexual preferences?

[mh1997]

If I had mod-points, I'd mod you up insightful. I didn't think someone would spot where I copied the review from so fast.

Re:Sexual preferences?

[Minwee]

Yes, they would have to have watched Hudson Hawk to do that. That narrows the field considerably.

Re:Sexual preferences?

[Jtheletter]

The search for good dialogue strays too far off the beaten path established by the social mores of the target market

I see what you've done there..... ;)

Re:Sexual preferences?

[caluml]

Yeah, I really liked it too. Quite surreal, funny, and the VHS copy I bought was purchased in Kazakhstan, so it has Russian subtitles, just to add to the weirdness. And when you tell people it has Bruce Willis in it, they're surprised.
Andi McDowell imitates a dolphin in it too.

Re:Sexual preferences?

[Minwee]

Is that any more surreal than a form of "entertainment" in which people get shot at or blown up every five minutes or so?

Probabilities

[dj245]

The researchers used this method to find how individuals on the IMDb privately rated films on Netflix, in the process possibly working out their political affiliation, sexual preferences and a number of other personal details"

This is a loaded statement. The most you can determine is that if a person likes movie A, B, C and D but hated E and F, there is a higher probability they are a guy. If they liked Z but didn't like X, there is a higher probability they might be a republican than not. You're still anonymous.

Unless, of course, you're one of the three people that liked "Glitter". Then I think they might have something on you.

Re:Probabilities

[Se7enLC]

I think they're on to something here. They cracked the anonymity by using the public movie ratings (and the dates those ratings were made) as a key. If the user has rated enough movies (especially some of the less-often-rated movies) you can uniquely identify which user they are. Once you know which user they are, you have now connected a username to the list of private ratings.

Now, they go one step too far to say that you can determine anything but movie preferences out of a movie rating list. Just because somebody liked or disliked brokeback mountain doesn't mean they are gay or straight, just like their opinion of michael moore movies doesn't give political affiliation.

It will tell you what movies they rented, though, and some people might not be happy having their movie-renting history publicly available.

Re:Probabilities

[Dare nMc]

one step too far to say that you can determine anything but movie preferences out of a movie rating list.

also your taking a aggregate of the household. So a household (will call them Chen'ys) had a gay kid, and the devil living in the same house with a Saint... good luck figuring out when the gay kid updates the queue, and when the Wife, or the Devil is at the keyboard.

Re:Probabilities

[Chapter80]

I think you're missing the point.

If you rate a handful of movies on ImDB, under the persona "MyNickname12345" and that can be traced to your personal MySpace page, you have made that choice. No problem.

If you then submit 100 movie ratings to Netflix, assuming that it is PRIVATE information that will not be linked back to you, and then Netflix releases the data to the public, now the 100 movies can be correlated to you, and your name can be revealed. Researchers have shown how PRIVATE DATA released to the public can be linked to already public information. PROBLEM!

Re:Probabilities

[coolGuyZak]

Some tech-savvy households may enable profiles on Netflix, enabling each person to track their likes & dislikes independently. (I did this for my GF, who has wildly disparate tastes from me). I'm not sure what effect that would have on the data. It'd certainly be neat if the scientists could differentiate between individual and multiple users using a particular profile.

Do what now?

[faloi]

It doesn't sound like the anonymity of the prize set was broken through any fault of NetFlix. It sounds like some sampling of users made the mistake of rating movies on a site where the info is publicly available, and a site where it's not. All they did was correlate the two.

So the lesson is, basically, don't post stuff that you don't want to be public to a website that makes it public, right? This is sounds roughly like blaming the DMV for figuring out a car owners likely political leanings by the bumper stickers on their car.

Re:Do what now?

[IBBoard]

Exactly - all they did was found that there was a correlation that might mean that the people are the same on IMDB and NetFlix. There's also the possibility that they're different people and that they just voted similar on different places.

Besides, this all relies on people voting for a) really obscure films so they can be easily identified and b) voting similarly or identically on lots of films so that they can get a better idea as to whether it is the same person based on them liking the same films the same amounts.

Just because two people from two different data sets both like (and are the only people in the data sets to like) lemon and custard jam as well as peanut butter with chips doesn't mean they're the same person, it just means they could be the same person and have similar tastes in obscure foods.

Re:Do what now?

[Peter Mork]

Exactly - all they did was found that there was a correlation that might mean that the people are the same on IMDB and NetFlix.

Caveat: I haven't had a chance to pore over the statistical calculations. However, the paper notes that their similarity measure was 38 standard deviations from the norm. Assuming the math is valid, this seems on par with a DNA test, which also provides a correlation. I wouldn't be so quick to dismiss the results until you can find a serious methodological problem.

Re:Do what now?

[JPMH]

Their lesson is that it can take surprising little public information to identify you.

For example, ratings on a scale of 1-5 for 2 movies, and a knowledge of when they were seen to within 14 days, was suffiecient to identify the complete data histories of 40% of the Netflix clients. As the authors say, that's the kind of information cooleagues give out every day around the water cooler.

Repeating the experiment with a knowledge of 8 movies, 6 hits in the database would be sufficient to identify the personal histories of 99% of clients included in the Netflix data.

Re:Do what now?

[roadkill_cr]

True, but in the real world, it's not as simple as that. There are cases of publicly available databases that you gave no permission to grant access to (for example, AOL's release of their search queries). There are other cases when a database has restricted access, but a person with access to it takes it and uses it in comparison with other databases available. Hackers are always a trouble; since some have gotten into such "secure" areas as the CIA and IRS, what's to keep them from potentially getting into any database?

The problem is one of privacy - in the worst case (or, for those who are cynical, common case) we have none. There's been some answers proposed to solve this. If you're interested, I'd start by reading the original paper on k-anonymity, which attempts to create privacy in a world where one can possibly have access to any database, ever. It can be found here: http://privacy.cs.cmu.edu/people/sweeney/kanonymity.html. (There are, of course, a multitude of other methods; k-anonymity is just a good starting point.)

Re:Do what now?

[IBBoard]

While yes, they did get a very perfect match on that record, the line about it is:

...our algorithm identified the records of two users the Netflix Prize dataset with eccentricities of around 28 and 15, respectively.

Granted they went for a small number of IMDB users due to their TOS, but that's still a tiny fraction. They mention finding a perfect match in IMDB and 1/8th of the NetFlix database towards the start of the report (although the sentence is a bit clunky and unclear). If that's their general accuracy then even if they can perfectly match some people (a statistical possibility) then they can't match enough to leave most people needing to worry.

Re:Do what now?

[arvindn]

"Besides, this all relies on people voting for a) really obscure films so they can be easily identified "

not true -- obscure films help a little bit but not too much. we put up a recent draft of our paper in which the dependence on obscure movies is much reduced.

"b) voting similarly or identically on lots of films so that they can get a better idea as to whether it is the same person based on them liking the same films the same amounts."

again not true at all. one of the main claims of our paper is that our method is tolerant to an INCREDIBLE amount of noise. we have the math to back this up.

--Arvind Narayanan

Re:Do what now?

[yali]

So the lesson is, basically, don't post stuff that you don't want to be public to a website that makes it public, right?

Nope, it's more complicated than that.

Suppose that you want to keep your political attitudes private -- for whatever reason, you decided it's nobody else's business. On IMDb, linked to your real identity, you only rate movies with non-political content, which you don't mind anybody knowing your opinion about. On Netflix, you believe that your ratings will be kept private, and you want to take advantage of their recommendations. So you rate all the same movies that you rated on IMDb, but you also post your ratings of Fahrenheit 9/11, The Corporation, etc. With the method described in this paper, somebody could potentially link your supposedly anonymized political ratings back to your real identity.

Anonymity broken by stupidity

[CastrTroy]

Seems like it was only broken because the identity of the people was posted somewhere else, along with the ratings. My only question is how they connected the rankings on Netflix, to the rankings on IMDB. Does Netflix take the liberty of submitting all the users rankings to IMDB for them, and also include their name with this data? If you just have anonymous dataset A, with anonymous dataset B, you could match up users from both and figure out which person in A is the same person in B, but you still wouldn't know who the person is. However, if you now have dataset B be not anonymous, then it's not too difficult to compare movie ratings and find out who the people are.

did it work?

[Speare]

The researchers used this method to find how individuals on the IMDb privately rated films on Netflix, in the process possibly working out their political affiliation, sexual preferences and a number of other personal details

{tongueincheek}Yeah, but the question is, will knowing those personal facts generate better movie recommendations?{/tongueincheek}

When there's a significant prize at stake, researchers can try all sorts of slimy tricks to win. (I'm not saying that's the motive behind this report, but there are many "researchers" going for the prize.) And when there's significant profits at stake, a corporation will damn-fire-certainly use whatever means they can use to maximize those profits, regardless of whether it might be "ethical."

Re:only a matter of time

[SatanicPuppy]

Perhaps if we're obscure and pretentious enough, no one will want to spy on us! Brillant!

The world changes. Learn to live with it.

How does this break anonymity?

[Anonymous Coward]

For those who haven't rated movies on IMDB, such as myself - and I imagine a large proportion of subscribers.

Data-mining and the actual problem

[Anonymous Coward]

There are two things going on here. One, many people are asking how you could identify any personal information about people based on their movie preferences. The answer is data-mining. Very sophisticated techniques exist to do things exactly like this, i.e. take a data set and find out about the people.

The second problem is that by deanonymizing the NetFlix data, you can start to cheat on the NetFlix prize. The requirement to win $1 million is that your recommendation engine is 10% better than the one they are currently using. However, if you can learn the exact preferences of some users in the dataset (i.e. by finding the rest of their ratings on IMDB) then you can hardcode that into your recommendation engine and get the recommendations for these users exactly right. This can boost your score even though your actual system is no better than the existing one. This is known as over-fitting to the data.

Finally, this paper is over a year old. Can we please have some new news?

Easy solution

[Thanshin]

Every time you feel the need to vote 10 in Glitter, also vote 10 to The Godfather.
Every time you cheer for Brokeback Mountain, also put a 10 in Huge Knockers MXII.
Every time you want to express your love for Dersu Uzala, vote a 10 in Spice World, with added commentaries.

That way, everybody will know you're a security conscious computer scientist. Or a squizophrenic moron.

requires another (partial)public revealing to work

[call -151]

The summary is somewhat misleading- the only accounts that can be identified are those that belong to people who also rate on IMBD and who have thus chosen to make at least some of their ratings public. If person X rates 1000 movies on Netflix and has made 20 or so ratings on IMDB publically available, then it is possible to infer with some small uncertainty which of the anonymized individuals in the NetFlix database they are. Thus you have possibly figured out their ratings of the other 980 movies they rated for Netflix but did not post on IMBD. Interesting, but not earth-shattering or a serious breach of privacy, I would say.

The world is not on fire

[puppetluva]

This is total hyperbole.

All they researchers are saying is that they can deduce some of your preferences based on your other preferences. Of COURSE you can do that, that was the whole point of the contest Netflix put up.

What they are _not_ saying is that they now know who you are, where you live, or anything uniquely identifying about you. So basically, you are still anonymous.

I'm starting to tire of news headlines that claim the world is on fire when someone actually just does something slightly derivative from the norm and thinks they are brilliant. The noise from these non-events mask actual brilliant achievements and make it seem that everyone is doing banal work.

Re:The world is not on fire

[Peter Mork]

All they researchers are saying is that they can deduce some of your preferences based on your other preferences.

The researchers are making a stronger claim. They are stating that based on actual public ratings (available from IMDB) they can generate actual private ratings published by Netflix under the guise of anonymity. As the paper notes, someone competing for the Netflix prize could use this data to improve the accuracy of their prediction algorithm. However, the point of this paper is to reveal that public ratings can be used to identify purportedly anonymous private ratings.

As a comparison, imagine if the public information consisted of the dates that various people went to the doctor for a yearly physical. This is hardly sensitive information. Now imagine that your insurance company provided a list of (id, date, diagnosis) records. Ostensibly, the id field is an arbitrary (anonymous) identifier. The paper shows that based on limited background knowledge (a handful of (date, 'physical exam') records), an attacker could reverse engineer your diagnosis history.

Re:The world is not on fire

[JPMH]

Othe the other hand, if somebody *already* knows who you are, the lesson is that it can take surprising little public information to identify your entire history of ratings at Netflix.

For example, the authors found for 40% of individuals, accurate ratings on a scale of 1-5 for only *two* random movies,together with a knowledge to within 14 days of when they were seen, would be sufficient to identify an individual in the dataset. As they comment, that's the kind of information cooleagues give out every day around the water cooler.

Repeating the experiment with a knowledge of 8 movies, 6 hits in the database would be sufficient to identify the personal histories of 99% of the people in that data.

What are you rating in IMDB vs Netflix

[SmallFurryCreature]

As far as I know in IMDB you are rating the overall quality of the movie, not I agree with it OR I want to see more like this.

One example, Shindlers list, great movie, do NOT want to see it again. Same with Grave of the fireflies. Some movies just ain't for multiple viewings. They are my "favorite movies I never want to see again".

On the other hand I got movies I can watch any day of the week, but that I would NEVER rate as highly. Cannonbal run is one such movie. It watch it far too often, but I wouldn't call it a good movie. You can always fine me ready for a Jacky Chan movie or a spagethi western.

Is the netflix rating system a "I liked this movie and want to see more like it" system or a "This movie was brilliant and I would highly recommend it too everyone else" type of rating system?

Granted some people get it confused, probably the same people that use the slashdot moderation system to silence views they don't like, but that only makes basing conclusions on user ratings even more problematic.

I can rate a movie highly even if I do not agree with it, simply because it is good. And I can rate a movie I really like to watch as crap simply because I know I like watching crap.

I don't like the godfather movies, I can see they are high quality, I just don't like them. So my rating them would be fairly high as for quality, but low for 'I want to see more like this'.

I thought that the netflix system was "I want to see more like this" based. Surely nobody is so stupid as to think a quality rating and a "i like this" rating system are the same? Or am I completly in the wrong in seeing a difference between the two? Am I insane in thinking that you can see a movie as being a great artwork and still not liking it or viceversa?

Re:What are you rating in IMDB vs Netflix

[xtracto]

One example, Shindlers list, great movie, do NOT want to see it again. Same with Grave of the fireflies. Some movies just ain't for multiple viewings. They are my "favorite movies I never want to see again".

Just out of curiosity, why don't you want to see those films again? both of them are really good films and although I would not see them every weekend (as for example Sin City), I enjoy watching them from time to time. The plot is interesting, the photography/drawing is nice and the screen writing is well done.

I find it difficult to understand your statement, "favorite movies I never want to see again", if you do not want to see them again, then you do not enjoy watching them... unless you dislike enjoyment and only watch films that make you cry or have a bad time (I would suggest you United 93... worst film I have seen in a looong long time... or Broeback Mountain, a 1 hour marlboro country ad).

I not not know about the netflix scoring algorithm but I have found criticker.com quite reliable for my tastes.

Am I insane in thinking that you can see a movie as being a great artwork and still not liking it or viceversa?
It might be akin to the "La Gioconda" painting. Everybody says it is the best piece of art of all the time, yet, after having watched it *twice* live in the Louvre I have yet to find something special about it (I prefer for example, paintings from Giovanni Paninni, which is relatively unknown)

Re:What are you rating in IMDB vs Netflix

[Danny Rathjens]

As far as I know in IMDB you are rating the overall quality of the movie, not I agree with it OR I want to see more like this.

No. You give people way too much credit if you think their ratings on public sites are that nuanced or objective. I think most people just rate things on how well they like it themselves. A significant portion seem to even just give 10s to anything they like, too.

I also find it amusing how the votes tend to congregate somewhere in the 3rd quartile a bit above average(e.g. 7 on a 1-10 scale) rather than 5.5 where it would be if people ranked things more fairly. (I wonder if this is associated with that effect where people always rank themselves above average despite evidence to the contrary, as well.)

Re:What are you rating in IMDB vs Netflix

[ps236]

> I also find it amusing how the votes tend to congregate somewhere in the 3rd quartile a bit above average(e.g. 7 on a 1-10 scale) rather than 5.5 where it would be if people ranked things more fairly

I'm not sure about that. People will tend to watch films they think/hope they will like. So, the ones where they think 'that'll be absolute poop' they won't bother watching, so, hopefully, won't bother rating.

So, people should rate fewer films as 'poop' than as 'great', because they select only the 'hopefully good' films to review.

If you forced people to go to see and review all films, even the ones where you have to drag them screaming through the door, then the average rating would almost certainly decrease considerably.

Re:requires another (partial)public revealing to w

[TubeSteak]

Interesting, but not earth-shattering or a serious breach of privacy, I would say. And who exactly are you to say so?
Because it isn't a Credit Card # or SSN it isn't serious?

A) Some people would rather go to jail or commit suicide than admit to something embarrassing they'd rather keep private. Privacy isn't (just) about hiding (illegal) things from the Government.

B) Demographic information is something you can never take back and can never change.
At least I can get a new credit card & SSN.

Re:This is a 'research' paper?

[nagora]

You're missing the point completely. Other people will be using "data mining" of this sort, and making serious decisions about whether you support terrorism, or are just generally not a "good citizen", and they won't be revealing their judgments to the public to let them know what might be going on.

TWW

Re:only a matter of time

[phobos13013]

Actually TFA seems to suggest that the more obscure and pretentious we are, the easier it is the track us. If we become homogeneous drones voting on the top 100 films, we are safe! Even so, I don't plan to become a homogeneous drone...

Re:This is a 'research' paper?

[amccaf1]

From TFA:

He did not like "Super Size Me" at all; perhaps this implies something about his physical size?

Or maybe he's a manager of a McDonalds. Or a part-time Ronald McDonald. Or...

Simple as you said, I do NOT enjoy watching them

[SmallFurryCreature]

The comment "favotire movie I never want to see again" is one I got from a review of Grave of the Fireflies that I just happened to totally agree with. Don't read the reviews, just watch it yourselve and if you are not into Anime just set that aside for the duration of the movie, then ask yourselve again, if you can understand that comment.

It is powerfull movie, like Shindlers List, but not a happy tale. I am not talking a tear jerker movie here, I am talking a "we will all burn in hell for this" movie. Tear jerkers I can take, Christmas in August is one. Sad tale, nicely told but ultimately human. It makes you sad, not sick of humanity.

Perhaps I am just too emotional about this kinda stuff, one reason might be that I grew up with halfunderstood tales of "that was were your great-uncle was picked up". When you realize just why your grandmother had 9 brothers and sisters yet you never met any. I got one aunt, my grand-parents had 3 kids, a starvation story like GotF hits a lot closer with a history like that. (The dutch hunger winter)

I enjoy all kinds of movies and would NOT have NOT watched these two, but that doesn't mean I want to see them again. There are some people who list Shindlers List as a feel good movie because it 'ends well'. I suppose you might see it that way, I don't.

I can regonize your statements that the photography is nice and the screen writing is well done, but the plot is intresting? To you it is a plot, to me it is a sickening part of history that I am far too close to.

Perhaps it is a bit like how Richard Pryor's monologue about the 200th celebration of the US was not exactly all that cheerfull.

Terry Pratchets Nanny Ogg describers at one point the difference between merry and mirth (or something like that) she describes how she was joyfull when her child was being born but she wasn't exactly chuckling at the time. Enjoying a movie and enjoying it are two different things, at least for me. I can't describe it any clearer.

More woe for HMRC then

[AmiMoJo]

None of the mainstream media picked up on it, but I remember thinking this sort of thing might be possible with the data lost by HMRC too. I bet Tesco would love to get their hands on it for planning where to put new stores and what to stock etc. Combined with their Clubcard database, of course.