First example you cite has been shown to be based on flawed statistics, i.e., the algorithm was shown not to produce biased results on the data. Bad things happen when journalists try to do statistical analysis.
Reference: Flores, Bechtel, Lowencamp; "False Positives, False Negatives, and False Analyses: A Rejoinder to âoeMachine Bias: Thereâ(TM)s Software Used Across the Country to Predict Future Criminals. And itâ(TM)s Biased Against Blacks.â", Federal Probation Journal, September 2016, You can find the article here: http://www.uscourts.gov/statis...
Slight tangent: The article cites the ProPublica study on the Northpointe software in which journalists (not statisticians) reported the software as biased. What they left out is that an independent study found this study showing bias to be wrong.
Source: Flores, Bechtel, Lowencamp; Federal Probation Journal, September 2016, "False Positives, False Negatives, and False Analyses: A Rejoinder to “Machine Bias: There’s Software Used Across the Country to Predict Future Criminals. And it’s Biased Against Blacks.”", URL http://www.uscourts.gov/statis...
In fact the ProPublica analysis was so wrong that the authors wrote: "It is noteworthy that the ProPublica code of ethics advises investigative journalists that "when in doubt, ask" numerous times. We feel that Larson et al.'s (2016) omissions and mistakes could have been avoided had they just asked. Perhaps they might have even asked...a criminologist? We certainly respect the mission of ProPublica, which is to "practice and promote investigative journalism in the public interest." However, we also feel that the journalists at ProPublica strayed from their own code of ethics in that they did not present the facts accurately, their presentation of the existing literature was incomplete, and they failed to "ask." While we aren’t inferring that they had an agenda in writing their story, we believe that they are better equipped to report the research news, rather than attempt to make the research news."
The authors of the ProPublica article are no longer with the organization, but this article shows up in any news article about AI bias. The fake story just doesn't want to die...
With all that said, I have some hopes that algorithms will help make truly race-blind decisions in criminal justice. It's easier to test them for bias than humans, and decisions are made in a consistent, repeatable manner.
With a lot of caveats, of course, but basically an individual has the right to 'an explanation of the decision reached after [algorithmic] assessment'. Described in http://fusion.kinja.com/eu-cit...
Interesting. Thanks for sharing. The decisions of linear models (like the ones used for sentencing, credit scores etc.) can indeed be explained. The credit bureaus sell explanations of the score and tips on improving the score as an extra service, for example. That doesn't mean that people will like the explanation... If the EU will indeed require explanations for algorithmic decisions, then models will be limited to simple linear models.
A justice system is about objectivity. It is not about punishing someone because you dislike them. Justice is blind, not a popularity contest.
What's more objective than using an algorithm for sentencing that is using data that has been shown to be predictive for recidivism? Indeed it's not about disliking somebody but about public safety and deterrence.
Frank Pasquale left out a couple of details in his opinion piece. First, these algorithms are only used in determining sentences, not to determine guilt. At that point in the trial guilt has been determined beyond a reasonable doubt. At the sentencing phase rules of evidence do not apply anymore and almost anything goes. That's why the prosecutor puts crying victims on the stand. There have been two supreme court cases (Malenchik vs Indiana and Loomis v. Wisconsin) that challenged the use of algorithms in sentencing and both upheld (in the later with some minor restrictions) that these algorithms can be used in sentencing. The conditions in general are that the algorithm has to be scientifically sound. That was the case in both cases that were challenged as there is existing peer-reviewed literature that examined the algorithms. Frank neglected to mention that. The secrecy of the algorithms is a consequence of patent and copyright law btw. The algorithms in these cases are a scoring function. Math is not (and should not) patentable or copyrightable. In this case the consequence is that the manufacturers only recourse is to keep it a trade secret. That could be solved better, but in my opinion people shouldn't get their hopes up that there's some exploitable loophole in the algorithm or something.
We can debate whether assessments (actuarial prediction instruments) should be used in sentencing or criminal justice. I’m very much in favor as it does reduce bias and leads to reproducible results. It’s much easier to control for biases in decision making with statistical methods than it is to control or fix bias in humans. Does anybody believe that human judgement is less biased? You can read up on the work of Paul Meehl who spend his lifetime showing that even simple assessment tests outperform the judgment of trained clinicians. Part of the sentencing is taking into consideration how likely the perpetrator is to commit a new offense. Humans suck at making predictions and estimating probabilities. This is no different in criminal justice.
Let me end this with pointing out some of the positive change that systems like this have brought: early release from incarceration. Low risk prisoners are more frequently released early (not just from overcrowded California prisons when ordered to do so by a federal judge), and then put on probation/parole. And work out well it did: http://time.com/4065359/califo... The expected crime wave from federally mandated early release didn't materialize. In my opinion thanks to these prediction models.
There are many things wrong with policing and criminal justice in the US, but the move to what’s generally referred to as “evidence based practices” (incl. actuarial prediction instruments) has been pretty positive. The great part is that both Dems and Reps are behind the idea of risk assessments so we might actually see some change for the better.
Re:Multivariate testing finally getting it's due..
on
Google Website Optimizer
·
· Score: 2, Interesting
Tell me about it. I'm currently running a little experiment which headline performs better and for shits and giggles threw in a "placebo"-headline (it is just a non-sense statement that has nothing to do with the product). The placebo outperformed the other headlines...
If their web-bot they have running in addition to their submission based system (google for "Turnitinbot") downloads papers I put on my webpage, but they won't allow me to search their database for free (like Google would), how would that be fair? Where's my gain from them using my work?
So it finally payed off that I forwarded all my spam to the SEC:-)
The current amount of stocks advertised by spam is human-manageable (see http://www.spamnation.info/stocks/ for a list of about 15 spamvertised symbols), and it will probably stay that way, because there are not that many companies that have the volume and the price such that spam will have an influence on the price; I think that is a very sensible solution for the spam-problem in that domain; I wish they could do something like that for medication-spam as well...
Patent holders must license or produce the product before they can sue anybody. That should make it a lot more difficult for patent trolls.
Prohibit people from suing private citizens for patent infringement - or at least limit the damages/legal costs for them.
Make with-holding prior-art from the examiner an offense; have the people sign an affidavit or something, and enforce it.
Have a higher burden of proof for the non-obviousness. Have the people that apply show to the examiner how their idea is different from what's out there.
No patents on business methods, algorithms, living organisms and such. This is ridiculous and got out of whack due to some messed up court ruling ("anything useful under the sun [] should be patentable"). Make a law to reserse said court ruling.
Maybe a public review period where prior art can be submitted to the examiner?
More examiners. I read somewhere that they have only about an hour or so to search for prior art, due to the small number of examiners the USPTO has.
I did one year of community service (Zivildienst) working in maintenance in a hospital back in Germany. We did stuff like cleaning up clogged up toilets and stuff like that. Most people would probably consider it common sense that you shouldn't flush big things like diapers or bandages down the toilet. This happened constantly. Not from paintients but with the chamber-pot cleaners the nurses were using. And that despite us explaining over and over again that they will clog. Things got better when we were (unfortunately) very busy, and couldn't clean them for a day or two. Just one episode where the friendly way didn't quite work...
Although there are probably unfriendly people in IT, I wonder if management will ever hold users accountable for viruses, spyware or whatnot installed after they were explicitly trained to avoid the problem. Just a thought.
While Sony is not a person, they are still liable. And there is an interesting twist in copyright law. IANAL, but I recall that the CEO can be held personally responsible for copyright infridgement of the company (that was in some recent/. article about copyright).
If the copyright-holder(s) of LAME and whatever other product they stole from actually files charges, then they are in really deep sh*t. Bringing a civil suit against Sony for copyright infridgement (and - as it seems to be industry standard - asking for $150.000 per infridgement) and maybe even filing criminal charges... That will wake all the copyright-nazis and DRM people up.
On a sidenote: TFA mentions that executives of companies can be held personally liable in cases of copyright infridgement (prison!). If that is true and if it can indeed be shown that Sony's CD contains LGPL code from LAME, then they have gotten themselfes into deeeeeep sh*t (although I doubt that anything that drastic would happen).
That sounds pretty old. Ever heard of Biometric encryption? The idea is to use a one-way hash on the biometrics, but also accounting for the fuzzy-ness in the reading. If the readings match, then the same hash comes out. Otherwise something random.
See here...
I recall that there was a patch for snort that was specifically designed to prevent people from breaking into other systems from a compromised honeypot machine. It did some good stuff like replacing NOP-slides with breakpoints etc. I don't have the URL handy, but this might help you with your intrusion prevention...
If I recall correctly Bittorrent will find out if a peer trying to cheat you and gives a corrupt piece. I don't think this is possible with their encoding scheme as it's entirely based on linear combinations. You won't be able to find out who cheated you and the entire file will be useless. That would be an easy way to disrupt people's downloads.
... the EU should have mandated the new edition to be cheaper than the Home XP? If I'd have to pay the same (or even a higher) price to get less, then guess what I would do. Without mandating it to be cheaper MS will probably resort to this kind of tactics to push their media-player.
Inflated clicks are not the only problem PPC concepts have lately. It's a pretty challenging problem to prevent click-fraud; open-proxies/botnets and so on make this even harder.
to educate users. If somebody signs up for a free-mail account (could work for ISPs in general as well), they are automatically send a couple of fake spams. If any link in the spam-emails is clicked, the account of the user will be closed (with an educational warning message). That will teach them...
Credit: Some MS guy I talked to. Unfortunately Hotmail-management was kinda opposed to that idea...
C't.
IMHO the best computer magazine out there, covering Windows, Linux and Mac. I also like their fair and balanced (no joke intended here) product tests.
First, turns out that the copy protection was lame (shift-key anyone?). So one reason for the lack of consumer backlash (if not just ignorance) might be that people figured that out...
On the other hand, what's dangerous is that the music industry might see this as the precedent to boost sales with copy protection. I mean, piracy (according to them) is hurting their buissiness, right? So it's logical from their point of view that a copy protected cd generates more sales and sales is the measure for the charts. This sets a bad precedent...
Re:What the law says and what's done in practice .
on
Canadian Privacy Act
·
· Score: 2, Informative
I know a bit more about it (though I'm not a lawyer):-)
Yes, unfortunately the law doesn't specify any about penalties. To the best of my knowledge the highest damages that has ever been awarded for a violation of the privacy rights was ~100k. Not bad, but that person was able to proof in court that he had suffered real monetary damages. Psychological distress doesn't count:-)
Courts have been relucatant with rewarding damages. For example, the phone company published a phone number of battered women's shelter by accident. They had to close the shelter, because they couldn't guarantee the safety anymore. They had to sell the house at a loss etc. and move elsewhere. The court awarded 15k in damages. That's a joke...
Another thing that the law describes is that you may only ask for the data you need. That has led to webmaster being "abgemahnt" (like a competitor complaining, costs you some money, but all without a court) for asking the name of newsletter-subscribers (email address would have been enough)... uh well...
But guess what... Some companies just moved their computing centers to chile, because they don't have privacy laws. They export the data, do the "illegal" cross linking in chile, and then re-import the data.
It's not that simple in practice. Getting damages from a court is nice, but German courts are a bit more realistic in awarding damages. What's easier is getting a court order to have them stop. While the law doesn't specify penalties/damages, violating a court order can get you in trouble...
I've seen discussion about this elsewhere a while ago. It's not the first time this topic pops up. Most webmasters sugested technical or legal solutions.
A popular suggestion was to trick the filter into blocking useful content to render the site unuseable. Those included, among javascript tricks (checking for the presence of a named image or something), having your navigation bar being an image map in banner size etc. I don't know if it's a good idea to exclude users/customers, but it's a way to draw the user's attention to the fact that there is a "problem" with their filter program. You can also probably include the word "advertisement" in some background colours into your links somehow...
Some other people considered to sue for copyright infridgement. IANAL, but the basic idea was to claim the modification of copyrighted work and redistribution (the filter is a proxy - so it takes a webpage and distributes a modified version). Don't know if that has any chance...
It seems that all those anoying, flashy banners finaly fire back...
Maybe that's just my paranoia, but assume for a moment that they implement it in a silly way (instead of mstnef), i.e. with an X-Do-not-forward header or something like that. Then every non-Outlook email program that doesn't have a full implementation of these extensions is a circumvention device (though not for the/sole purpose/ as the DMCA reuqires).
Maybe we should implement that first for Mutt and then we sue MS for circumventing our "security system":-)
Does anybody know any details of how MS's system works? Is it some special attachment or something like that?
Slashdot Mirror
First example you cite has been shown to be based on flawed statistics, i.e., the algorithm was shown not to produce biased results on the data. Bad things happen when journalists try to do statistical analysis.
Reference: Flores, Bechtel, Lowencamp; "False Positives, False Negatives, and False Analyses: A Rejoinder to âoeMachine Bias: Thereâ(TM)s Software Used Across the Country to Predict Future Criminals. And itâ(TM)s Biased Against Blacks.â", Federal Probation Journal, September 2016, You can find the article here: http://www.uscourts.gov/statis...
Slight tangent: The article cites the ProPublica study on the Northpointe software in which journalists (not statisticians) reported the software as biased. What they left out is that an independent study found this study showing bias to be wrong.
Source: Flores, Bechtel, Lowencamp; Federal Probation Journal, September 2016, "False Positives, False Negatives, and False Analyses: A Rejoinder to “Machine Bias: There’s Software Used Across the Country to Predict Future Criminals. And it’s Biased Against Blacks.”", URL http://www.uscourts.gov/statis...
In fact the ProPublica analysis was so wrong that the authors wrote: "It is noteworthy that the ProPublica code of ethics advises investigative journalists that "when in doubt, ask" numerous times. We feel that Larson et al.'s (2016) omissions and mistakes could have been avoided had they just asked. Perhaps they might have even asked...a criminologist? We certainly respect the mission of ProPublica, which is to "practice and promote investigative journalism in the public interest." However, we also feel that the journalists at ProPublica strayed from their own code of ethics in that they did not present the facts accurately, their presentation of the existing literature was incomplete, and they failed to "ask." While we aren’t inferring that they had an agenda in writing their story, we believe that they are better equipped to report the research news, rather than attempt to make the research news."
The authors of the ProPublica article are no longer with the organization, but this article shows up in any news article about AI bias. The fake story just doesn't want to die...
With all that said, I have some hopes that algorithms will help make truly race-blind decisions in criminal justice. It's easier to test them for bias than humans, and decisions are made in a consistent, repeatable manner.
With a lot of caveats, of course, but basically an individual has the right to 'an explanation of the decision reached after [algorithmic] assessment'. Described in http://fusion.kinja.com/eu-cit...
Interesting. Thanks for sharing. The decisions of linear models (like the ones used for sentencing, credit scores etc.) can indeed be explained. The credit bureaus sell explanations of the score and tips on improving the score as an extra service, for example. That doesn't mean that people will like the explanation... If the EU will indeed require explanations for algorithmic decisions, then models will be limited to simple linear models.
A justice system is about objectivity. It is not about punishing someone because you dislike them. Justice is blind, not a popularity contest.
What's more objective than using an algorithm for sentencing that is using data that has been shown to be predictive for recidivism? Indeed it's not about disliking somebody but about public safety and deterrence.
Frank Pasquale left out a couple of details in his opinion piece. First, these algorithms are only used in determining sentences, not to determine guilt. At that point in the trial guilt has been determined beyond a reasonable doubt. At the sentencing phase rules of evidence do not apply anymore and almost anything goes. That's why the prosecutor puts crying victims on the stand. There have been two supreme court cases (Malenchik vs Indiana and Loomis v. Wisconsin) that challenged the use of algorithms in sentencing and both upheld (in the later with some minor restrictions) that these algorithms can be used in sentencing. The conditions in general are that the algorithm has to be scientifically sound. That was the case in both cases that were challenged as there is existing peer-reviewed literature that examined the algorithms. Frank neglected to mention that. The secrecy of the algorithms is a consequence of patent and copyright law btw. The algorithms in these cases are a scoring function. Math is not (and should not) patentable or copyrightable. In this case the consequence is that the manufacturers only recourse is to keep it a trade secret. That could be solved better, but in my opinion people shouldn't get their hopes up that there's some exploitable loophole in the algorithm or something.
We can debate whether assessments (actuarial prediction instruments) should be used in sentencing or criminal justice. I’m very much in favor as it does reduce bias and leads to reproducible results. It’s much easier to control for biases in decision making with statistical methods than it is to control or fix bias in humans. Does anybody believe that human judgement is less biased? You can read up on the work of Paul Meehl who spend his lifetime showing that even simple assessment tests outperform the judgment of trained clinicians. Part of the sentencing is taking into consideration how likely the perpetrator is to commit a new offense. Humans suck at making predictions and estimating probabilities. This is no different in criminal justice.
Let me end this with pointing out some of the positive change that systems like this have brought: early release from incarceration. Low risk prisoners are more frequently released early (not just from overcrowded California prisons when ordered to do so by a federal judge), and then put on probation/parole. And work out well it did: http://time.com/4065359/califo... The expected crime wave from federally mandated early release didn't materialize. In my opinion thanks to these prediction models.
There are many things wrong with policing and criminal justice in the US, but the move to what’s generally referred to as “evidence based practices” (incl. actuarial prediction instruments) has been pretty positive. The great part is that both Dems and Reps are behind the idea of risk assessments so we might actually see some change for the better.
Tell me about it. I'm currently running a little experiment which headline performs better and for shits and giggles threw in a "placebo"-headline (it is just a non-sense statement that has nothing to do with the product). The placebo outperformed the other headlines ...
If their web-bot they have running in addition to their submission based system (google for "Turnitinbot") downloads papers I put on my webpage, but they won't allow me to search their database for free (like Google would), how would that be fair? Where's my gain from them using my work?
So it finally payed off that I forwarded all my spam to the SEC :-)
The current amount of stocks advertised by spam is human-manageable (see http://www.spamnation.info/stocks/ for a list of about 15 spamvertised symbols), and it will probably stay that way, because there are not that many companies that have the volume and the price such that spam will have an influence on the price; I think that is a very sensible solution for the spam-problem in that domain; I wish they could do something like that for medication-spam as well...
Patent holders must license or produce the product before they can sue anybody. That should make it a lot more difficult for patent trolls.
Prohibit people from suing private citizens for patent infringement - or at least limit the damages/legal costs for them.
Make with-holding prior-art from the examiner an offense; have the people sign an affidavit or something, and enforce it.
Have a higher burden of proof for the non-obviousness. Have the people that apply show to the examiner how their idea is different from what's out there.
No patents on business methods, algorithms, living organisms and such. This is ridiculous and got out of whack due to some messed up court ruling ("anything useful under the sun [] should be patentable"). Make a law to reserse said court ruling.
Maybe a public review period where prior art can be submitted to the examiner?
More examiners. I read somewhere that they have only about an hour or so to search for prior art, due to the small number of examiners the USPTO has.
I did one year of community service (Zivildienst) working in maintenance in a hospital back in Germany. We did stuff like cleaning up clogged up toilets and stuff like that. Most people would probably consider it common sense that you shouldn't flush big things like diapers or bandages down the toilet. This happened constantly. Not from paintients but with the chamber-pot cleaners the nurses were using. And that despite us explaining over and over again that they will clog. Things got better when we were (unfortunately) very busy, and couldn't clean them for a day or two. Just one episode where the friendly way didn't quite work...
Although there are probably unfriendly people in IT, I wonder if management will ever hold users accountable for viruses, spyware or whatnot installed after they were explicitly trained to avoid the problem. Just a thought.
While Sony is not a person, they are still liable. And there is an interesting twist in copyright law. IANAL, but I recall that the CEO can be held personally responsible for copyright infridgement of the company (that was in some recent /. article about copyright).
If the copyright-holder(s) of LAME and whatever other product they stole from actually files charges, then they are in really deep sh*t. Bringing a civil suit against Sony for copyright infridgement (and - as it seems to be industry standard - asking for $150.000 per infridgement) and maybe even filing criminal charges... That will wake all the copyright-nazis and DRM people up.
I don't expect this to happen, though.
On a sidenote:
TFA mentions that executives of companies can be held personally liable in cases of copyright infridgement (prison!). If that is true and if it can indeed be shown that Sony's CD contains LGPL code from LAME, then they have gotten themselfes into deeeeeep sh*t (although I doubt that anything that drastic would happen).
There are two linux-distros for the Linksys WRT54G that are meant to help with setting up hotspots:
EWRT Linux http://www.portless.net/menu/ewrt/ and the hotspot-zone project http://sourceforge.net/projects/hotspot-zone/. Both use nocat as the captive portal, the later offers radius authentication patches for nocat.
I recall that there was a patch for snort that was specifically designed to prevent people from breaking into other systems from a compromised honeypot machine. It did some good stuff like replacing NOP-slides with breakpoints etc. I don't have the URL handy, but this might help you with your intrusion prevention...
If I recall correctly Bittorrent will find out if a peer trying to cheat you and gives a corrupt piece. I don't think this is possible with their encoding scheme as it's entirely based on linear combinations. You won't be able to find out who cheated you and the entire file will be useless. That would be an easy way to disrupt people's downloads.
... the EU should have mandated the new edition to be cheaper than the Home XP? If I'd have to pay the same (or even a higher) price to get less, then guess what I would do. Without mandating it to be cheaper MS will probably resort to this kind of tactics to push their media-player.
Inflated clicks are not the only problem PPC concepts have lately. It's a pretty challenging problem to prevent click-fraud; open-proxies/botnets and so on make this even harder.
A bunch of interesting links:
to educate users. If somebody signs up for a free-mail account (could work for ISPs in general as well), they are automatically send a couple of fake spams. If any link in the spam-emails is clicked, the account of the user will be closed (with an educational warning message). That will teach them...
Credit: Some MS guy I talked to. Unfortunately Hotmail-management was kinda opposed to that idea...
C't. IMHO the best computer magazine out there, covering Windows, Linux and Mac. I also like their fair and balanced (no joke intended here) product tests.
First, turns out that the copy protection was lame (shift-key anyone?). So one reason for the lack of consumer backlash (if not just ignorance) might be that people figured that out...
On the other hand, what's dangerous is that the music industry might see this as the precedent to boost sales with copy protection. I mean, piracy (according to them) is hurting their buissiness, right? So it's logical from their point of view that a copy protected cd generates more sales and sales is the measure for the charts. This sets a bad precedent...
I know a bit more about it (though I'm not a lawyer) :-)
:-)
Yes, unfortunately the law doesn't specify any about penalties. To the best of my knowledge the highest damages that has ever been awarded for a violation of the privacy rights was ~100k. Not bad, but that person was able to proof in court that he had suffered real monetary damages. Psychological distress doesn't count
Courts have been relucatant with rewarding damages. For example, the phone company published a phone number of battered women's shelter by accident. They had to close the shelter, because they couldn't guarantee the safety anymore. They had to sell the house at a loss etc. and move elsewhere. The court awarded 15k in damages. That's a joke...
Another thing that the law describes is that you may only ask for the data you need. That has led to webmaster being "abgemahnt" (like a competitor complaining, costs you some money, but all without a court) for asking the name of newsletter-subscribers (email address would have been enough)... uh well...
But guess what... Some companies just moved their computing centers to chile, because they don't have privacy laws. They export the data, do the "illegal" cross linking in chile, and then re-import the data.
It's not that simple in practice. Getting damages from a court is nice, but German courts are a bit more realistic in awarding damages. What's easier is getting a court order to have them stop. While the law doesn't specify penalties/damages, violating a court order can get you in trouble...
I've seen discussion about this elsewhere a while ago. It's not the first time this topic pops up.
Most webmasters sugested technical or legal solutions.
A popular suggestion was to trick the filter into blocking useful content to render the site unuseable. Those included, among javascript tricks (checking for the presence of a named image or something), having your navigation bar being an image map in banner size etc. I don't know if it's a good idea to exclude users/customers, but it's a way to draw the user's attention to the fact that there is a "problem" with their filter program. You can also probably include the word "advertisement" in some background colours into your links somehow...
Some other people considered to sue for copyright infridgement. IANAL, but the basic idea was to claim the modification of copyrighted work and redistribution (the filter is a proxy - so it takes a webpage and distributes a modified version). Don't know if that has any chance...
It seems that all those anoying, flashy banners finaly fire back...
Maybe that's just my paranoia, but assume for a moment that they implement it in a silly way (instead of mstnef), i.e. with an X-Do-not-forward header or something like that. Then every non-Outlook email program that doesn't have a full implementation of these extensions is a circumvention device (though not for the /sole purpose/ as the DMCA reuqires).
:-)
Maybe we should implement that first for Mutt and then we sue MS for circumventing our "security system"
Does anybody know any details of how MS's system works? Is it some special attachment or something like that?