Posts Tagged ‘Weapons of Math Destruction’

The Electoral College Needs to Go

May 17, 2018

This post is based on Cathy O’Neil’s informative book, “Weapons of Math Destruction.” The penultimate chapter in the book shows how weapons of math destruction are ruining our elections. It is only recently that Facebook and Cambridge Analytics have be found to employ users data for nefarious purposes. Nevertheless Dr. O’Neil’s book was published in 2016. To summarize the chapter, weapons of math destruction are distorting if not destroying our elections. Actually the most informative and most important part of the chapter is found in a footnote at the end:

“At the federal level, this problem could be greatly alleviated by abolishing the Electoral College system. It’s the winner-take-all mathematics from state to state that delivers so much power to a relative handful of voters. It’s as if in politics, as in economics, we have a privileged 1 percent. And the money from the financial 1 percent underwrites the micro targeting to secure the votes of the political 1 percent. Without the Electoral College, by contrast, every vote would be worth exactly the same. That would be a step toward democracy. “

Readers of the healthy memory blog should realize that the Electoral College is an injustice that has been addressed in previous healthy memory blog posts (13 to be exact). Just recently, the Electoral College, not the popular vote, produced Presidents with adverse effects. One resulted in a war in Iraq that was justified by nonexistent weapons of mass destruction. And most recently, the most ill-suited person for the presidency became president, contrary to the popular vote.

The justification for the Electoral College was the fear that ill-informed voters might elect someone who was unsuitable for the office. If there ever was a candidate unsuitable for the office, that candidate was Donald Trump. It was the duty of the Electoral College to deny him the presidency, a duty they failed. So, the Electoral College needs to be disbanded and never reassembled.

© Douglas Griffith and healthymemory.wordpress.com, 2018. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Douglas Griffith and healthymemory.wordpress.com with appropriate and specific direction to the original content.

Advertisements

Broken Windows Policing

May 16, 2018

This post is based on Cathy O’Neil’s informative book, “Weapons of Math Destruction.” The title of this post should be familiar to anyone who has viewed the Blue Bloods television series. It advanced Broken Windows Policing as justification for the policies they pursued to prevent serious crimes. The justification of this policy has been an article of faith since 1982, when a criminologist named George Kelling teamed up with a public policy expert, James Q. Wilson to write an article in the “Atlantic Monthly” on so-called broken-windows policing. According to Dr. O’Neil, “The idea was that low-level crimes and misdemeanors created an atmosphere of disorder in a neighborhood. This scared law-abiding citizens aware. The dark and empty streets they left behind were breeding grounds for serious crimes. The antidote was for society to resist the spread of disorder. This included fixing broken windows cleaning up graffiti-covered subway cars, and taking steps to discourage nuisance crimes. This thinking led in the 1990s to zero-tolerance campaigns most famously in New York City. Cops would arrest people for jumping subway turnstiles. They’d apprehend people caught sharing a single joint and rumble them around the city in a paddy wagon for hours before eventually booking them.”

There were dramatic campaigns for violent crimes. The zero-tolerance campaign was credited for reducing violent crime. Others disagreed citing the fallacy of “post hoc, propter hoc” (after this, therefore because of this) and other possibilities, ranging from the falling rates of crack cocaine addiction to the booming 1990s economy. Regardless, the zero-tolerance movement gained broad support, and the criminal justice system sent millions of mostly young minority males meant to prison, many of them for minor offenses.

Dr. O’Neil continues, “But zero tolerance actually had very little to do with Kelling and Wilson’s “broken-windows” thesis. Their case focused on what appeared to be a successful policing initiative in Newark, New Jersey. Cops who walked the beat there, according to the program, were supposed to be highly tolerant. Their job was to adjust to the neighborhood’s own standards of order and to help uphold them. Standards varied from one part of the city to another. In one neighborhood it might mean that drunks had to keep their bottles in bags and avoid major streets but that side streets were okay. Addicts could sit on stoops but not lie down. The idea was only to make sure the standards didn’t fall. The cops, in this scheme, were helping a neighborhood maintain its own order but not imposing their own.”

On the basis of this and other data, Dr. O’Neil comes to the conclusion, “that we criminalize poverty, believing all the while that our tools are not only scientific, but fair.” Dr. O’Neil asks, “What if police looked for different kinds of crimes?” That may sound counterintuitive, because most of us, including the police, view crime as a pyramid. At the top is homicide. It’s followed by rape and assault, which are more common, then shoplifting, petty fraud, and even parking violations, which happen all the time. Minimizing violent crime, most would agree, is and should be a central part of a police force’s mission.”

Dr. O’Neil asks an interesting question. What if we looked at the crimes carried out by the rich? “In the 2000s, the kings of finance threw themselves a lavish party. They lied, they bet billions against their own customers, they committed fraud and paid off rating agencies. Enormous crimes were committed there, and the result devastated the global economy for the best part of five years. Millions of people lost their homes, jobs, and health care.”

She continues,”We have every reason to believe that more such crimes are reoccurring in finance right now. If we’ve learned anything, it’s that the driving goal of the finance world is to make a huge profit, the bigger the better, and that anything resembling self-regulation is worthless. Thanks largely to the industry’s wealth and powerful lobbies, finance is underpoliced.”

Two Especially Troubling Problems

May 15, 2018

One of these problems is found in the Chapter “Propaganda Machine: Online Advertising in Dr. Cathy O’Neil’s book “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy”. Advertising is legitimate, but predatory advertising is certainly not. In predatory advertising weapons of math destruction are used to identify likely subjects to be exploited. Not all, but some for-profit colleges were built and grew through weapons of math destruction. People who were identified as being in need of education or training were preyed upon and sold expensive on-line courses, that were not likely to pay off in jobs or any sort of advancement.

HM learned a new word reading Dr. Kathy O’Neil’s book “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy”. That word was clopening. This is when an employee works late one night to close the store or cafe and then returns a few hours later, before dawn, to open it. Having the same employee closing and opening, or clopening, can make logistical sense for a company, but it leads to sleep-deprived workers and crazy schedules. Weapons of math destruction can identify optimal schedules for the company, but they also need to take into account the welfare of the employee. Scheduling can place the employee’s health in jeopardy along with the employee’s family life.

Laws are clearly needed here. As for the predatory advertisers marketing on-line courses, they should be closed down and fined. Unfortunately, the Consumer Financial Protection Bureau that was policing this problem has been shut down. Companies and businesses need to be held responsible for the health and welfare of their employees.

© Douglas Griffith and healthymemory.wordpress.com, 2018. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Douglas Griffith and healthymemory.wordpress.com with appropriate and specific direction to the original content.

The General Problem of Proxies

May 14, 2018

This general problem of proxies is fairly ubiquitous as outlined in Dr. Cathy O’Neil’s book “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy”. Remember that proxies are variables used to compensate for the actual variables for which data are unavailable. The Chapter “Ineligible to Serve” addresses problems proxies can create in getting a job. Once on the job proxies can make it more difficult to hold the job. This is described in the chapter, “Sweating Bullets: on the Job.” Proxies also cause problems in getting credit, which is described in the chapter “Collateral Damage: Landing Credit.” Similarly proxies present problems in getting insurance described in the chapter, “No Safe Zone: Getting Insurance.”

So the effects of Weapons of Math Destruction are ubiquitous. People need to be aware of when they might be being screwed by these weapons. So “Weapons of Math Destruction” needs to be generally read.

Indeed, there are reasons why these weapons are being used, but care must be taken to reduce or eliminate the destruction. It is not only the individuals being evaluated who need to be aware, but also the businesses and agencies using them. They should be aware of their shortcomings and the need for eliminating these shortcomings when possible. These models need to be made transparent, so the proxies can be identified, and the possibility of misclassifications can be addressed.

There is also a chapter titled “The Targeted Citizen,” but since that topic is so much in the news about Facebook and the interference of Russia in the presidential election, that will not be addressed here.

Ranking Colleges

May 13, 2018

This post is based on Dr. Cathy O’Neil’s book “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy”.

In 1983 the newsmagazine “U.S. News & World Report” decided it would evaluate 1,800 colleges and universities throughout the United States and rank them for excellence. Had they honestly considered if they could accurately do this they could have saved the country and the countries’ colleges and universities from anxiety and confusion. But they were not honest and proceeded to build the magazine’s reputation and fortune.

How could one do this? One could conduct a national survey and have individuals rate the schools in terms of prestige. This could be done validly. But to rate them in terms of excellence? How is excellence defined? Would it be the satisfaction of recent graduates? Would it be the satisfaction of graduates further down the course of life?

The healthy memory blog has made the point in previous posts that depending on what a student wants to learn and what career the student wants to pursue should be primary factors in choosing a college. All colleges, even the most prestigious ones, differ in what they have to offer. And what about the cost-effectiveness of colleges? This is probably the most important factor for the majority of students. One can pay through the nose to attend a prestigious college, but what is the benefit for the cost incurred?

The magazine picked proxies that seemed to correlate with success. They looked at SAT scores, student-teacher ratios, and acceptance rates. They analyzed the percentage of incoming freshmen who made it to the sophomore year and the percentage of those who graduated. They calculated the percentage of living alumni who contributed money to their alma mater, surmising that if they gave a college money there was a good chance they appreciated the education there. Three-quarters of the ranking would be produced by an algorithm, an opinion formalized in code, that incorporated these proxies. In the other quarter they would factor in the subjective views of college officials throughout the country.

HM regards this procedure pretty much as ad hoc selection with no external validation. However, Dr. O’Neil is more charitable writing, “U.S. News first data-driven ranking came out in 1988, and the results seemed sensible. However, as the rankings grew into a national standard, a vicious feedback loop materialized. The trouble was that the rankings were self-reinforcing.” So if a college was rated poorly in “U.S. News,” its reputation would suffer, and conditions would deteriorate. Top students would avoid it, as would top professors. Alumni would howl and cut back on contributions. The ranking would go down further. Dr. O’Neil concludes that the ranking was destiny.

Everyone was acting foolishly. In fact, this was a jury-rigged methodology that provided a proxy estimate of a school’s prestige. ‘U.S. News” should have discontinued the survey. Universities should have disclaimed the methodology and the ratings. Instead, they played the game and took actions just to improve their ratings. Read the book to learn the gory details.

Dr. O’Neil notes that when you create a model from proxies, it is far simpler to game it. This is because proxies are easier to manipulate than the complicated reality they represent. This is a common problem with big data and weapons of math destruction.

© Douglas Griffith and healthymemory.wordpress.com, 2018. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Douglas Griffith and healthymemory.wordpress.com with appropriate and specific direction to the original content.

Finance and Big Data

May 12, 2018

This post is based on Dr. Cathy O’Neil’s book “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy”. Dr. O’Neil was originally applying her mathematical knowledge and skills in finance. In 2008 there was a catastrophic market crash. Although weapons of math destruction did not solely cause the financial crash, they definitely contributed to it. So Dr. O’Neil moved from finance to Big Data where her skills were readily transferable.

She writes, “In fact, I saw all kinds of parallels between finance and Big Data. Both industries gobble up the same pool of talent, much of it from elite universities like MIT, Princeton, or Stanford. These new hires are ravenous for success and have been focused on external metrics—like SAT scores and college admissions—their entire lives. Whether in finance or tech, the message they’ve received is that a they will be rich, that they will run the world. Their productivity indicates that they’re on the right track, and it translates into dollars. This leads to the fallacious conclusion that whatever they’re doing to bring in more money is good. It ‘adds value.’ Otherwise, why would the market reward it?”

She continues, “In both of these industries, the real world, with all of its messiness sits apart. The inclination is to replace people with data trails, turning them into more effective shoppers, voters, or workers to optimize some objective. This is easy to do, and to justify, when success comes back as an anonymous score and when the people affected remain ever bit as abstract as the numbers dancing across the screen.”

She worried about the separation between technical models and real people and about the moral repercussions of the separation. She saw the same pattern emerging in Big Data that she’d witnessed in finance: a false sense of security was leading to widespread use of imperfect models, self-serving definitions of success, and the growing feedback loops.

She continued working in Big Data. She writes that the her journey to disillusionment was more or less complete, and the misuse of mathematics was accelerating. She started a blog on this problem and in spite of almost daily blogging she barely kept up with all the ways she was hearing of people being manipulated, controlled, and intimidated by algorithms. It began with teachers working under inappropriate value-added models (read the book to learn about this), then the LSI-R risk model, and and continued from there. She quit her job to investigate full time the issue leading to this book.

Three Kinds of Models

May 11, 2018

This post is based on Dr. Cathy O’Neil’s book “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy”. Many of us likely develop predictive models, but remain unaware what we are doing. So Dr. O’Neil describes an internal intuitive model she uses in planning family meals. She has a model of everyone’s appetite. She knows that one of her sons loves chicken (but hates hamburgers), while another will eat only pasta (with extra grated parmesan cheese). She also has to take into account that people’s appetites vary from day to day, so a change can catch her internal model by surprise. In addition to the information she has about her family, she knows the ingredients she has on hand or knows are available, plus her own energy, time, and ambition. The output is how and what she decides to cook. She evaluated the success of a meal by how satisfied her family seems at the end of it, how much they’ve eaten, and how healthy the food was. Seeing how well it is received and how much of it is enjoyed allows her to update her model for the next time she cooks. These updates and adjustments make it what is called a “dynamic model.”
Her model is a good model as long as she restricts it to her family. The technical term for this limitation is that it doesn’t scale. It will not work with larger or different families.

Examples of the best models are those used by professional baseball teams. There are an enormous number of variables that can be used to predict a teams performance. Moreover, these models allow the prediction of the performance of the team when different players are added or subtracted. The measure this model is designed to predict is the number of wins. Wins provides the variable that it used to predict and improve the models.

Recidivism models are used to predict the likelihood that a prisoner, after being released from prison will return to criminal behavior and end up back in jail. One of the more popular models is the Level of Service Inventory-Revised (LSI-R). It includes a lengthy questionnaire for the prisoner to fill out. One of the questions—“How many prior convictions have you had?” is highly relevant to the risk of recidivism. Others are also clearly related. For example “What part did others play in the offense? What part did drugs and alcohol play?”

Other questions are more problematic. For example a question about the first time they ever were involved with the police. For a white subject the only incident to report might be the one that brought him to prison. However, young black males are likely to have been stopped by police dozens of times, even when they’ve done nothing wrong. A 2013 study by the New York Civil Liberties Union found the while black and Latino males between the ages of fourteen and twenty-four make up only on 4.7% of the cities population, but accounted for 40.6% of the stop-and-frisk checks by police. More than 90% of those stopped were innocent. Some of the others might have been drinking underage or carrying a joint. And unlike most rich kids, they got in trouble for it. So if early “involvement” with police signals recidivism, poor people and racial minorities look far riskier.

Although statistical systems like the LSI-R are effective in gauging recidivism risk, or at least more accurate than a judge’s random guess, we find ourselves descending into a pernicious WMD feedback loop. A person who scores as “high risk” is likely to be unemployed and to come from a neighborhood where many of his friends and family have had run-ins with the law. Dr. O’Neil writes, “Thanks in part to the resulting high score on the evaluation, he gets a longer sentence, locking him away for more years in a prison where he’s surrounded by criminals, which raises the likelihood that he’ll return to prison. If he commits another crime, the recidivism model can claim another success. But in fact the model contributes to a toxic situation and helps to sustain it. That’s a signature quality of a WMD.

This risk and the value of the LSR-R could be tested. There could be two groups. A control group would be administered the questionnaire. Another group would be administered a modified version of the questionnaire that did not include responses that would tip the race of the individual. The participants could be tracked over time. If the modified version of the questionnaire actually resulted in the a lower rate of recidivism, then the original questionnaire could be identified as harmful, not only to the respondent, but also to society that was increasing recidivism rather than reducing it.

© Douglas Griffith and healthymemory.wordpress.com, 2018. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Douglas Griffith and healthymemory.wordpress.com with appropriate and specific direction to the original content.