Deep-Fried Data • Hacker News Board

Re: Deep-Fried Data

#101

Post by wrsh07 » Mon, Oct 03, 2016, 12:47 PM UTC

yummyfajitas wrote: post #97

pcmaffey wrote: post #59

Machine learning does not have less bias than human researchers. It is simply magnified at scale. And that scale is exactly the state of the internet. There is so much data available to study and understand, that we absolutely need better tools, like machine learning or whatever we want to call it, to help us keep up. Shit's moving faster than our human perception can handle, especially for those who didn't grow up w…

This is simply not true. Most algorithms can and will correct for biases in their inputs. See this (somewhat technical) article where I go into explicit (simulations in numpy) levels of detail: https://www.chrisstucchio.com/blog/2016/alien_intelligences_... The best analogy I've come up with for the non-technical is that algorithms are like humans trying to draw inferences on octopus society. Some octopi might have b…

This is very optimistic. There are well known and documented cases of ml algorithm bias and its cause [1].

And it's not surprising that data itself contains some biases from the humans creating it. Suppose police are asking machine learning where more crime is committed - there will be a feedback loop. Where are they currently making more arrests? If they spend more time there, the bias will be exaggerated.

The op correctly gauges how we should be cautious. Your post, I'm afraid, is misleading at best.

[1] https://www.google.com/amp/s/www.technologyreview.com/s/6017...

Re: Deep-Fried Data

#102

Post by b2m9 » Mon, Oct 03, 2016, 1:08 PM UTC

rubidium wrote: post #98

iyn wrote: post #77

Earlier quoted context omitted.

I agree with that, but I have a small suggestion until these points are reality: you can consider adding/enabling SSL/TLS for your blog. Thanks! P.S. I really like your posts and your tweets are hilarious, please don't ever stop.

Ok, this tweaks my curiosity. Why would one put SSL on a static personal/blog type website (assuming one doesn't care about the google penalty)?

For example, ISPs are not able to crawl your traffic if it's via HTTPS. I've worked on data sets gathered by major ISPs and it's scary how much they know about their users (especially if they also have a mobile phone with the same company). ISPs use such intelligence for personalised marketing (either for their own product catalogue or 3rd parties)

Re: Deep-Fried Data

#103

Post by yummyfajitas » Mon, Oct 03, 2016, 1:42 PM UTC

wrsh07 wrote: post #101

yummyfajitas wrote: post #97

Earlier quoted context omitted.

This is simply not true. Most algorithms can and will correct for biases in their inputs. See this (somewhat technical) article where I go into explicit (simulations in numpy) levels of detail: https://www.chrisstucchio.com/blog/2016/alien_intelligences_... The best analogy I've come up with for the non-technical is that algorithms are like humans trying to draw inferences on octopus society. Some octopi might have b…

This is very optimistic. There are well known and documented cases of ml algorithm bias and its cause [1]. And it's not surprising that data itself contains some biases from the humans creating it. Suppose police are asking machine learning where more crime is committed - there will be a feedback loop. Where are they currently making more arrests? If they spend more time there, the bias will be exaggerated. The op co…

Of course data contains biases. But again, please read the article I linked; algorithms will have a tendency to correct that bias.

The examples in the article you link to are not algorithmic bias at all. They consist of:

1) Humans at Facebook manipulating trending results.

2) Google's keyword algorithm (accurately) reflecting the fact that people with black names are more likely to have arrest records.

Lets distinguish "bias" from "accurately learning things you wish it wouldn't learn" or "accurately learning things you wish weren't true."

None of what I'm saying is remotely controversial. If I told you statistics could detect and correct bias in a mobile phone compass, you'd just think "cool stats bro". Is this article remotely controversial? https://www.chrisstucchio.com/blog/2016/bayesian_calibration...

The specific feedback loop you describe - variable detection probability => variable # of detections - can be directly mitigated. For a non-controversial example drawn from sensor networks (sensors report events with a delayed eraction, the longer you wait the more events you detect), see here: https://www.chrisstucchio.com/blog/2016/delayed_reactions.ht...

(You can find similar examples all over the place. I just link to the ones I wrote because they spring immediately to mind.)

In a compass, a sensor network, adtech or other quant finance, the idea that machine learning can fix biased inputs is not remotely controversial. The concept that statistics suddenly stops working to fix racism is just silly anthropomorphism.

Re: Deep-Fried Data

#104

Post by mbrock » Mon, Oct 03, 2016, 1:50 PM UTC

rubidium wrote: post #98

iyn wrote: post #77

Earlier quoted context omitted.

I agree with that, but I have a small suggestion until these points are reality: you can consider adding/enabling SSL/TLS for your blog. Thanks! P.S. I really like your posts and your tweets are hilarious, please don't ever stop.

Ok, this tweaks my curiosity. Why would one put SSL on a static personal/blog type website (assuming one doesn't care about the google penalty)?

Otherwise network operators can sniff, alter, and generally fuck with the integrity of the site and its users.

Re: Deep-Fried Data

#105

Post by erichocean » Mon, Oct 03, 2016, 2:13 PM UTC

idlewords wrote: post #69

erichocean wrote: post #45

> Publish your texts as text. Let the images be images. Put them behind URLs and then commit to keeping them there. I sounds like he's saying ephemeral content is worthless and should be shunned. I, and hundreds of millions of others, disagree. You want a bland, awful, boring society? Easy: make everything you do stick around forever—like a promise. And then watch the world self-police as the lifeblood drains out of…

The audience for this talk was people with very large collections they're bringing online. I was trying to encourage them to avoid exotic formats, custom plugins, custom software (shudder) when they put this material online, and make them web accessible. For example, here is three quarters of a PETABYTE of historical American newspapers: http://chroniclingamerica.loc.gov

Okay, that makes sense. I 100% agree that bringing collections online in exotic formats is a terrible idea.

Re: Deep-Fried Data

#106

Post by yummyfajitas » Mon, Oct 03, 2016, 2:14 PM UTC

ohthehugemanate wrote: post #100

pcmaffey wrote: post #59

Machine learning does not have less bias than human researchers. It is simply magnified at scale. And that scale is exactly the state of the internet. There is so much data available to study and understand, that we absolutely need better tools, like machine learning or whatever we want to call it, to help us keep up. Shit's moving faster than our human perception can handle, especially for those who didn't grow up w…

Bias is a pattern-generation process. Machine learning is a pattern-recognition process. Any bias on the part of the (human) data collection, or the (human) training program author, gets spit out as a "pattern", because it is one. The problem is that it gives the illusion of a bias in reality. My go-to example is machine learning police enforcement direction, often used as a counter to racially biased policing. This…

Of course, in reality this bias can be corrected for. I don't know if specific crime-stat software does it, but it's certainly doable. Here's an example where I solve literally the same problem (better measurement => more events detected) in a different scenario. Is it remotely controversial that I can do this for a sensor network?

https://www.chrisstucchio.com/blog/2016/delayed_reactions.ht...

You might also be interested to know that a variety of studies have shown that policing is not particularly biased. Arrest statistics and the like correspond pretty well with NCVS and similar crime victim surveys.

http://slatestarcodex.com/2014/11/25/race-and-justice-much-m...

Re: Deep-Fried Data

#107

Post by wrsh07 » Mon, Oct 03, 2016, 2:38 PM UTC

yummyfajitas wrote: post #103

wrsh07 wrote: post #101

Earlier quoted context omitted.

This is very optimistic. There are well known and documented cases of ml algorithm bias and its cause [1]. And it's not surprising that data itself contains some biases from the humans creating it. Suppose police are asking machine learning where more crime is committed - there will be a feedback loop. Where are they currently making more arrests? If they spend more time there, the bias will be exaggerated. The op co…

Of course data contains biases. But again, please read the article I linked; algorithms will have a tendency to correct that bias. The examples in the article you link to are not algorithmic bias at all. They consist of: 1) Humans at Facebook manipulating trending results. 2) Google's keyword algorithm (accurately) reflecting the fact that people with black names are more likely to have arrest records. Lets distingui…

Aha - I think I see our miscommunication. When you say bias you mean statistical bias.

Yes, machine learning is able to correct for that kind of bias - 538's polls forecast is a good example of that.

But you don't get to redefine racial bias to be something innocuous. Yes, black names are more likely to have arrest records, but that "fact" is super misleading [1].

Finally, you're talking past me. I'm not saying that statistics is broken. I'm saying that we should be especially mindful of the OPs point when they say this:

> So what’s your data being fried in? These algorithms train on large collections that you know nothing about. Sites like Google operate on a scale hundreds of times bigger than anything in the humanities. Any irregularities in that training data end up infused into in the classifier.

I think the OP author also has a related post about the kind of bias I'm talking about: http://idlewords.com/talks/sase_panel.htm

[1]: http://www.huffingtonpost.com/kim-farbota/black-crime-rates-...

Re: Deep-Fried Data

#108

Post by Asooka » Mon, Oct 03, 2016, 2:49 PM UTC

pimlottc wrote: post #19

> Many [programmers] work jobs that are intellectually stimulating, but ultimately leave nothing behind. There is a large population of technical people who would enjoy contributing to something lasting. This hits pretty close to home.

But if you built things that lasted, there would be less work for the next generation. That's one of the reasons I don't really worry about the constant reinventing going on. It may be unfortunate from a theoretical perspective, but is an economic godsend for each new generation.

Re: Deep-Fried Data

#109

Post by yummyfajitas » Mon, Oct 03, 2016, 3:59 PM UTC

wrsh07 wrote: post #107

yummyfajitas wrote: post #103

Earlier quoted context omitted.

Of course data contains biases. But again, please read the article I linked; algorithms will have a tendency to correct that bias. The examples in the article you link to are not algorithmic bias at all. They consist of: 1) Humans at Facebook manipulating trending results. 2) Google's keyword algorithm (accurately) reflecting the fact that people with black names are more likely to have arrest records. Lets distingui…

Aha - I think I see our miscommunication. When you say bias you mean statistical bias. Yes, machine learning is able to correct for that kind of bias - 538's polls forecast is a good example of that. But you don't get to redefine racial bias to be something innocuous. Yes, black names are more likely to have arrest records, but that "fact" is super misleading [1]. Finally, you're talking past me. I'm not saying that…

Without getting into a dispute about the definition of "bias", I'm saying that algorithms can accurately measure reality even if input(x=white, all else equal) != input(x=black, all else equal).

You are saying that algorithms are accurately measuring a reality you wish were different. I don't disagree with this.

The right thing to do is to actually answer unpleasant moral questions like "if blacks are 4x more likely to be dangerous criminals, what should we do about it?" But I guess overloading the word "bias" is a nice substitute for clearly thinking things through.

Re: Deep-Fried Data

#110

Post by mrow84 » Mon, Oct 03, 2016, 4:08 PM UTC

yummyfajitas wrote: post #97

pcmaffey wrote: post #59

Machine learning does not have less bias than human researchers. It is simply magnified at scale. And that scale is exactly the state of the internet. There is so much data available to study and understand, that we absolutely need better tools, like machine learning or whatever we want to call it, to help us keep up. Shit's moving faster than our human perception can handle, especially for those who didn't grow up w…

This is simply not true. Most algorithms can and will correct for biases in their inputs. See this (somewhat technical) article where I go into explicit (simulations in numpy) levels of detail: https://www.chrisstucchio.com/blog/2016/alien_intelligences_... The best analogy I've come up with for the non-technical is that algorithms are like humans trying to draw inferences on octopus society. Some octopi might have b…

To correct biased measurements (in a careful way) you need

1. Enough knowledge about the structure of the bias to be able to devise a model for it.

2. Some measurements from which to fit the model, with errors that are uncorrelated with the errors in your original data.

These things are not always easy to obtain, even in relatively mundane settings. It is also a distinctly non-automatic procedure - it requires someone to decide that a bias exists, to model it, obtain the relevant data, and fit the bias correction model, all before they can begin to obtain unbiased (or probably just less-biased) measurements.