How Cambridge Analytica’s Facebook targeting model really worked • Hacker News Board

Re: How Cambridge Analytica’s Facebook targeting model really worked

#141

Post by superioritycplx » Sat, Mar 31, 2018, 6:36 AM UTC

vuln wrote: post #77

frozenlettuce wrote: post #66

Earlier quoted context omitted.

It became a "problem" because it helped Trump win.

This is exactly it. At least it stops the news from droning on and on about Russia. I thought Clinton spent large amounts of money on data and the Democrats admitted the data was bad or at least that was their excuse. How much did CA pay for this data? I still find it crazy that Trump campaign spent 30% of what Hillary did and still won. The Russians used 100k$ worth of ads to sway the election. This stuff doesn't t…

According to the VP of facebook ads the majority of ads were purchased AFTER the election: https://gizmodo.com/trump-eagerly-jumps-on-facebook-ad-vps-a...

Re: How Cambridge Analytica’s Facebook targeting model really worked

#142

Post by cosmic_ape » Sat, Mar 31, 2018, 7:21 AM UTC

Besides the data, its interesting how the actual targeting was performed.

Does facebook provide an option to show a particular given ad to a particular given user? Or is it possible to select a group of people with a given set of likes? How fine-grained is facebook's audience selection mechanism for ads?

Or was the targeting performed by creating fake groups, befriending people?

Re: How Cambridge Analytica’s Facebook targeting model really worked

#143

Post by stevenicr » Sat, Mar 31, 2018, 8:34 AM UTC

golemotron wrote: post #56

RpFLCL wrote: post #34

Earlier quoted context omitted.

One example I can provide is of gun control topics. If you understand someone's mentality on the subject you can decide if they see: 1) An ad with someone breaking into a home and the homeowner defending themselves with a firearm (sell insurance?) 2) A grandfather and grandson on a hunting trip (hunting supplies?) 3) Or maybe gun violence hotline with powerful images. The people seeing these ads are under the assumpt…

> The people seeing these ads are under the assumption that everyone else sees them, not that it's specifically targeted at their personality type. How long will that be true? Do people make that assumption about search results?

I think most people do assume the same with search results. How many do you think assume that the ads on the TV they see could be different than what the neighbor is seeing when watching the same channel with the same cable company? I think a lot of people assume that others see the same news and the way people act you'd think they assume that others see the same things in their newsfeed/timeline / facebook thing - and wonder how others could have a different view.

Even when I explain how ads can be different, I don't think people really want to believe it, or understand it, and they certainly do not realize the power of these targeting abilities..

Re: How Cambridge Analytica’s Facebook targeting model really worked

#144

Post by mrkgnao » Sat, Mar 31, 2018, 8:37 AM UTC

stochastic_monk wrote: post #84

darawk wrote: post #68

Earlier quoted context omitted.

> when strictly speaking the raw data has indeed been deleted after being used to create a derivative work that can for all important purposes be used to recreate the original? To be precise, you almost certainly cannot use this data to recreate anything remotely resembling the original dataset. This type of dimensionality reduction would throw away enormous volumes of data. There is no meaningful sense in which you…

That's only accurate in the sense that because an LSTM's hidden layer is much smaller in dimension than the data on which it is trained, there is less information in it. However, it concisely represents a manifold in a much larger dimensional space and effectively captures most of the information in it. It may be (and is) lossy, but don't underestimate the expressive power of a deep neural network.

Alternatively, "If I take a FLAC you own, make a 320kbps MP3 from it, and store it on my laptop, am I still in possession of any IP belonging to you?"

Re: How Cambridge Analytica’s Facebook targeting model really worked

#145

Post by stevenicr » Sat, Mar 31, 2018, 8:42 AM UTC

darawk wrote: post #68

etiam wrote: post #41

Spoiler warning. Article punchline ahead. "The whole point of a dimension reduction model is to mathematically represent the data in simpler form. It’s as if Cambridge Analytica took a very high-resolution photograph, resized it to be smaller, and then deleted the original. The photo still exists — and as long as Cambridge Analytica’s models exist, the data effectively does too." That's an eloquent piece of explanati…

> when strictly speaking the raw data has indeed been deleted after being used to create a derivative work that can for all important purposes be used to recreate the original? To be precise, you almost certainly cannot use this data to recreate anything remotely resembling the original dataset. This type of dimensionality reduction would throw away enormous volumes of data. There is no meaningful sense in which you…

" cannot use this data to recreate anything remotely resembling the original dataset."

This by itself may be mostly true perhaps - and many of the comments get into ways of playing with this dataset to make it better, I don't have experience with those methods, but,

what I have not seen anyone mention, if you have this dumbed down dataset, the original is gone.. you can still combine with other data sets that are either public or previously created and likely fine tune;

dumbed down set + public voter records + public arrest records + previous whatever records - sort, match, what's left over.

and pretty much recreate what you needed from the original, maybe not 100%, but I would guess you could get really close.

Re: How Cambridge Analytica’s Facebook targeting model really worked

#146

Post by StringyBob » Sat, Mar 31, 2018, 9:10 AM UTC

mrkgnao wrote: post #144

stochastic_monk wrote: post #84

Earlier quoted context omitted.

That's only accurate in the sense that because an LSTM's hidden layer is much smaller in dimension than the data on which it is trained, there is less information in it. However, it concisely represents a manifold in a much larger dimensional space and effectively captures most of the information in it. It may be (and is) lossy, but don't underestimate the expressive power of a deep neural network.

Alternatively, "If I take a FLAC you own, make a 320kbps MP3 from it, and store it on my laptop, am I still in possession of any IP belonging to you?"

I like that analogy. I'll make it more tenuous with - "I took a copy of your album collection without your permission, ripped them to MP3, played them so much everyone is sick of them. but you've still got all the original CD's you don't even use, so no problem right?"

On this tangent, IP ownership for deep learning models is interesting - how to you prove (in court) someone has/hasn't copied model/stolen a training set? If you fed someone else's training/model into your system, how easy is it to prove? Will we see the equivalent of map 'trap streets' in trained CNN models?

Which led me to: https://medium.com/@dtunkelang/the-end-of-intellectual-prope...

Re: How Cambridge Analytica’s Facebook targeting model really worked

#147

Post by etiam » Sat, Mar 31, 2018, 9:26 AM UTC

darawk wrote: post #140

stochastic_monk wrote: post #123

Earlier quoted context omitted.

Outside of the fact that they have identities for all of the people whose data they acquired, yes, it would be harder to reconstruct individual people with it than PCA because of the direct interpretability of its data.

They claim to have deleted that data. If they haven't deleted the data, then of course it's still an invasion of privacy. But the ML model really has nothing to do with it.

I think the ML model has a lot to do with it in this case. One of the arguments I expect to see is that "Oh, no! We removed all the data. It's gone. I mean, that was only a few hundred megabytes per person anyway, but we just calculate a few thousand numbers from it and save in our system, then delete the data. That's less data per person than is needed to show a short cute cat GIF. What harm could we possibly do with that?"

Re: How Cambridge Analytica’s Facebook targeting model really worked

#148

Post by makomk » Sat, Mar 31, 2018, 9:47 AM UTC

mrkgnao wrote: post #144

stochastic_monk wrote: post #84

Earlier quoted context omitted.

That's only accurate in the sense that because an LSTM's hidden layer is much smaller in dimension than the data on which it is trained, there is less information in it. However, it concisely represents a manifold in a much larger dimensional space and effectively captures most of the information in it. It may be (and is) lossy, but don't underestimate the expressive power of a deep neural network.

Alternatively, "If I take a FLAC you own, make a 320kbps MP3 from it, and store it on my laptop, am I still in possession of any IP belonging to you?"

I think a better analogy might be "If I take a few hundred thousand MP3s and come up with a clever way to reduce each to a short representation of its genre, mood, tempo, etc that can be used to identify similar music, then throw away all the original MP3s, am I still in possession of the original music". The whole point is to turn the individual data into broad, general categorisations that are easier to handle because they contain much less information. Remember, they're using this for ad targeting, and the reason they're doing it is so they can target broad groups of people rather than having to manually go through and target ads at each individual one by one.

Re: How Cambridge Analytica’s Facebook targeting model really worked

#149

Post by jobigoud » Sat, Mar 31, 2018, 9:56 AM UTC

etiam wrote: post #41

Spoiler warning. Article punchline ahead. "The whole point of a dimension reduction model is to mathematically represent the data in simpler form. It’s as if Cambridge Analytica took a very high-resolution photograph, resized it to be smaller, and then deleted the original. The photo still exists — and as long as Cambridge Analytica’s models exist, the data effectively does too." That's an eloquent piece of explanati…

This analogy immediately reminds me that with enough high res / low res pairs we can rebuild a high resolution image from its low res version with fairly good results. Wonder if the same could be done here.

Re: How Cambridge Analytica’s Facebook targeting model really worked

#150

Post by YeGoblynQueenne » Sat, Mar 31, 2018, 10:03 AM UTC

stochastic_monk wrote: post #84

darawk wrote: post #68

Earlier quoted context omitted.

> when strictly speaking the raw data has indeed been deleted after being used to create a derivative work that can for all important purposes be used to recreate the original? To be precise, you almost certainly cannot use this data to recreate anything remotely resembling the original dataset. This type of dimensionality reduction would throw away enormous volumes of data. There is no meaningful sense in which you…

That's only accurate in the sense that because an LSTM's hidden layer is much smaller in dimension than the data on which it is trained, there is less information in it. However, it concisely represents a manifold in a much larger dimensional space and effectively captures most of the information in it. It may be (and is) lossy, but don't underestimate the expressive power of a deep neural network.

Where do neural nets come into this? The Koscinski-Stillwell-Graepel paper talks about using the reduced-dimensionality data with logistic regression.