How Cambridge Analytica’s Facebook targeting model really worked

Post by darawk » Sat, Mar 31, 2018, 9:07 PM UTC

etiam wrote: post #153

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…

> 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 can reconstruct the data from it. First off, I think that's wrong. The idea is after all to keep the information that will result in the smallest error compared to the origina…

Post by fjsolwmv » Sat, Mar 31, 2018, 11:14 PM UTC

alexcnwy wrote: post #106

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.

You're throwing out buzzwords instead of addressing the response. It's dimensionality reduction. You cannot recover the original object. It's like using a shadow to reconstruct the face of the person casting the shadow. Note this has nothing to do with the expressive power of a deep neural network. You are by definition trying to throw away noisy aspects of the data and generalize a lower dimensional manifold from a…

Post by fjsolwmv » Sun, Apr 01, 2018, 12:08 AM UTC

daodedickinson wrote: post #57

lordnacho wrote: post #10

Finally, I was waiting for someone to talk about the model itself. It makes sense that SVD or something like it (PCA, co-occurrence, etc) would be used. But I also wonder what exactly you are going to do with the predictions. What exactly do you show to someone to make them more likely to go and vote if they are inclined to vote your way, or make them stay at home otherwise? Is there evidence that whatever you're sho…

Door to door canvassers these days carry devices that tell you what topics to bring up and what topics not to bring up at a certain address, even distinguishing between individuals at an address; some are told to demand a husband let them talk to the wife, for example.

Post by rhizome » Sun, Apr 01, 2018, 12:48 AM UTC

stochastic_monk wrote: post #95

salty_biscuits wrote: post #88

Earlier quoted context omitted.

Only if the "true" data actually lives in a lower dimensional manifold and the data acurrately can encode it with low noise. I doubt anyone can tell who you will vote for depending on which cat videos you liked, no matter how magic your regressor.

I do think that the most significant components of a personality will likely be targetable with a relatively low nuclear norm. And, for example, where someone's proclivity on the exploration/exploitation spectrum, if you will, (IE, how strongly do they respond to fear-based messaging) falls is probably quite predictable from a spectrum of likes. Cat pictures may be less informative, but not all of these people clicke…

Post by mtimjones » Sun, Apr 01, 2018, 1:32 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…