mathbabe

I was involved in an interesting discussion the other day with other data scientists on the mistake people make when they describe a “racist algorithm”. Their point, which I largely agreed with, is that algorithms are simply mirroring back to us what we’ve fed them as training data, and in that sense they are no more racist than any other mirror. And yes, it’s a complicated mirror, but it’s still just a mirror.

This issue came up specifically because there was a recent Mic.com story about how, if you google image search “professional hairstyles for work,” you’ll get this:

but if you google image search “unprofessional hairstyles for work” you’ll instead get this:

This is problematic, but it’s also clearly not the intention of the Google engineering team, or the google image search algorithm, to be racist. It is instead a reflection of what we as a community have presented to that algorithm as “training data.” So in that sense we should blame ourselves, not the algorithm. The algorithm isn’t (intentionally) racist, because it’s not intentionally anything.

And although that’s true, it’s also dodging some other truth about how we talk about AI and algorithms in our society (and since we don’t differentiate appropriately between AI and algorithms, I’ll use them interchangeably).

Namely, we anthropomorphize AI all the time. Here’s a screenshot of what I got when I google image searched the phrase “AI”:

Out of the above images, only a couple of them do not have some reference to human brains or bodies.

In other words, we are marketing AI as if it’s human. And since we do that, we are treating it and reacting to it as quasi-humans would. That means when it seems racist, we’re going to say the AI is racist. And I think that, all things considered, it’s fair to do this, even though there’s no intention there.

Speaking of intention and blame, I am of the mind that, even though I do not suspect any Google employee of making their algorithms prone to this kind of problem, I still think they should have an internal team that’s on the look-out for this kind of thing and address it. Just as, as a parent, I am constantly on the look-out for my kids getting the wrong ideas about racism or other prejudices; I correct their mistakes. And I know I’m anthropomorphizing the google algorithms when I talk about them like children, but what can I say, I am a sucker for marketing.

I’ve been going to Alt Banking every week for almost 5 years. Here’s what I’ve learned:

There are lots of ways to break up the big banks.

First, there are ways to create incentives for banks to get smaller by themeselves. For example, we could impose more capital requirements on bigger banks than smaller ones, or more regulations, or we could say that banks beyond a specific size couldn’t engage in certain kinds of behavior, or trade certain kinds of derivatives, or we could impose taxes on those trades that are heavier for larger banks. Progressive taxes that max out at 100% profit when the bank is as big as Bank of America.

Next, we could outlaw huge banks. We could do this by simply defining a legal limit to the size of a bank, or its geographic scope, or the amount of risk it carries. We could even say that its connections to the other financial institutions has to be adequately uncomplicated that, in an event of bankruptcy, we could let it fail and it wouldn’t be a biggie. We could give banks 4 years to get compliant.

Or we could go nuts and say that banks are no longer able to “create money” at all, which is to say we could put an end to fractional reserve lending. Or, we could just do that for big banks, where reserve requirements would get larger as banks get larger. We could change accounting laws around banking to make it a lot harder for them to hide risk. We could make it illegal for them to trade derivatives, or impose a new version of Glass-Steagall to make certain things illegal. We could impose all sorts of laws that would blow them out of the water permanently.

Or we could go the other way entirely, and nationalize one of the big banks, or all of them, and let them remain big but treat them like utilities, and make them super duper boring, and pay no bonuses and have the CEO get paid something like $200K.

And I’m not saying what specific we should do – my heart’s probably closest to imposing high capital requirements and limited trading for big banks so they’ll make themselves smaller – but in any case, there are plenty of things we could do that there is no political appetite for. Once again, there’s no shortage of possible plans.

What we do have a shortage of, though, is political will. That’s why we don’t talk about this very much, and we might not have talked about it at all had Sanders not entered the race.

Which is why, when I hear people complaining about Sanders not have a specific enough plan for breaking up the big banks, I think it’s hogwash. What Sanders has, which is sorely needed, in fact is the crucial missing ingredient to any discussion around big banks, is the political will to do something. Once the political will is there, the details can be sorted out by people who think about this stuff all the time.

This is a guest post by Maxwell Feiner, a New York City high school kid interested in math. Maxwell and I have been having fun math conversations on Friday afternoons for a while now, and I’ve been impressed by the tools he uses, so I asked him to write up a description of them for mathbabe.

When I am doing math in my spare time there are three tools that I use heavily to aid in the process. These three being Desmos.com and Wolframalhpa.com for the aid in solving problems, and Brilliant.org for finding great problems to solve.

---

 

Brilliant

Brilliant is a fantastic site for users to obtain unique challenging problems, as well as to post solutions to problems posed by others. Think of it as a social site for math. Most of the problems are math, but there are some physics and chemistry problems as well. I like the problems a lot because they require insight beyond what is taught in school classes. They normally cannot be solved using one formula or pre-learned method, but instead require deeper thought and a combination of different concepts in order to be solved. Signing up is required to use the site. It is free. Some sample problems are shown below.

---

 

Desmos

Desmos is a great, free, and interactive online graphing calculator. It is simple to use, but at the same time very powerful. Besides just graphing equations, the user can put in adjustable values of variables and watch how the graph changes as the variables do. For example, the user could enter the equation y=a*sin(bx+c) and create adjustable values for a,b, and c, then see how changing them affects the graph of the function.

 

To change the values of the variables, the user can use either sliders (shown), manually input values, or put in a set of numbers or range, such as c=[1,2], where two graphs will be displayed for c=1 and c=2, or  c=[1,2,…10](also shown), where 10 graphs will be displayed, for every integer from 1 to 10. The sliders can also be used to make animations by continually and smoothly incrementing the value of one of the variables. In the graph pictured below I also used the restrictions capability, allowing me to only show a certain part of the graph.

Another cool feature is movable points, which is demonstrated in two of my own graphs which have links to them below. There are so many features that it would be very hard to explain them all here, so I have provided a link to some of their tutorials on their website and a few examples of some interactive graphs I created. You can create an account and publish your work too.

Here’s a tutorial, and here are some examples of my work:

1. Tangents to a parabola,
2. the equation of a square, and
3. construction of a sine wave

---

 

WolframAlpha

WolframAlpha is a powerful computational knowledge engine that can be used for many purposes, both math and non-math related. One thing it is particularly helpful for is graphing 3-D equations. Just type in an equation,1-D,2-D, or 3-D and it will be graphed. Once on the site, some interesting queries to try include the ones pictured below.

Other ones to try (just copy and paste into WolframAlpha or click on the links below)

• 
• 
• area under from 1 to 2
• rotation of 45 degrees
• 
• Parabola vs hyperbola
• (Population of China)/(Population of U.S.)

This is just barely scratching the surface of what can be done with the powerful tool.

---

 

Conclusion

After working with these three tools for a while, I, and many others have benefitted so much from using them. If you are looking for aids in solving problems, or more problems to solve, I highly recommend checking them out. 

1. Manpons:

The only thing missing from this ad is something analogous to the “And now it has 5 blades. Because you couldn’t possibly do with only 4 blades” one-upmanship of the men’s razor industry.

2. I am going to this:

The full name of the conference is The Color of Surveillance: Georgetown Law Conference to Explore Racial Bias of Government Monitoring, and I’m looking forward to it. Anyone worried about the dystopian future of government surveillance should learn about what’s happening right now to poor minority neighborhoods.

3. I’m not going to this:

Here’s a ridiculous quote from this conference which, typically, conflates what millennials “like” with what they’ve had to put up with because they were born into a world of enormous student debt and terrible job security (h/t Ernie Davis):

Game technologies are becoming increasingly popular in the workplace since they appeal to the millennial generation who have grown up playing video and computer games together with using mobile devices. But this is much more than fun! Serious games can generate up to millions of data points that can then be fed into machine-learning algorithms to help employers make smart HR decisions to win the war for talent.

4. I’m Worried About Self-driving Cars

Because won’t it encourage enormously wasteful use of cars? If I can go to sleep while I’m in my car and it’s driving to Lexington, Virginia, what will stop me from visiting my buddy Aaron every weekend?

And for that matter, why even be in the car while it’s going somewhere? I can send my car to do errands I don’t want to do, or deliver packages I don’t want to bother sending from the post office.

Last night I had drinks with someone who knows a ton about the Apple vs. FBI case. He explained to me the following:

• The way the FBI eventually figured out a way into the San Bernadino shooter’s phone was extremely involved and expensive, involving things like shaving tiny pieces of hardware apart without dropping anything or exposing anything to too much heat.
• This is a good thing, because that expensive process is extremely hard to scale.
• Also, there was no legal precedent created.
• Moreover, Apple has been making iPhones increasingly secure by default, for example with default encrypted iCloud data in more recent version of its operating system.
• Which means that in a couple of years, most people using iPhones will be pretty well protected from even expensive FBI searches, again as long as there’s no legal requirement to create backdoors.

This story is interesting, but it still leaves me extremely unsatisfied. In particular, I’ve really gotten riled up by stupid media stories that “Apple won”.

I’ve maintained for a while that this story isn’t a story about Apple at all, because Apple is not accountable to the public in any real way; Tim Cook could change his mind tomorrow about whether to care about consumer security and we wouldn’t be able to do anything about it.

I think I have to amend that claim somewhat, though. Because what’s really happening is that Apple, or rather Tim Cook, is pushing through his vision of consumer protection, knowing that there will be very little the U.S. government, or any other government for that matter, will be able to do about it on a technical level, unless they’re willing to make iPhones illegal altogether.

That’s not without precedent. For example, there are some radio scanners that are illegal in the U.S. and other countries. But it’d be hard to imagine what the public’s response would be to being told that they can no longer buy iPhones.

So, the way I see it, it’s Apple vs. everyone, and Apple is feeling pretty good about its chances.

And look, I happen to agree with Apple this time. But it’s a screwed up and tenuous situation, and it’s deeply anti-democratic. We haven’t actually had the urgently needed conversation about whether Americans have the right to encrypted communication. Instead, we’re relying on a private company to make de facto policy for our benefit. What?

Here’s what I’d like to see: a real conversation about what Americans are entitled to. It’s a conversation that Obama started a couple of weeks ago at SXSW:

if the government can’t get in, then everybody is walking around with a Swiss Bank account in their pocket. There has to be some concession to the need to be able to get into that information somehow.

I’ll start. Obama’s comparing the individual’s desire for privacy with a Swiss Bank account is a smear tactic on the one hand – we’re trying to avoid taxes or something, which smacks of the tired line “don’t worry if you have nothing to hide” – and it’s disingenuous on the other hand – acting as if all information is equivalent, when we know that the government may claim access to our financial information, for tax purposes, but should never have access to our love letters. And since both kinds of information is stored on our phone, I think right there we have a pretty great argument explaining why our phones are nothing like Swiss Bank accounts.

Here’s what I’d like to see. A nuanced discussion about what types of data the government should have access to and under what circumstances, where the government has to make its case and the public gets to weigh in, since we care about terrorism too.

1. The life and death of a metric

There’s a problem, or at least potentially a problem. Someone figures out how to measure the problem. The measurement isn’t perfect, and everyone admits that, but nevertheless nobody argues against using it, since knowing something is better than knowing nothing.

That metric is used a few times, and people get used to hearing it, and they mostly forget what its limitations were. Moreover, they start assigning blame to it. Instead of seeing a “bad score” as something the indicates a need for more resources and support, it becomes a moral failing: take responsibility for your terrible score and do something about it.

Stakes get high. People are measured, judged, and rewarded or punished based on their score. They start focusing on improving their score at all costs, and the small imperfections of the scoring system in the first place are magnified and distorted. Cheating happens too.

Before long, it’s all about the score, at least until dissenting voices point out that all this focus on the score hasn’t actually addressed the initial problem. In fact, it’s gotten worse over time. The focus on this metric is given up by some, held on to by others who have found other uses for it, and everyone starts looking around for a new metric to solve the initial problem.

2. Example: education reform

A few decades ago we decided to look into the international competitiveness of our nation’s children. We developed tests to see how much people in different states and different schools knew about certain things. This wasn’t a perfect process, to be sure, since the curriculums varied from state to state and school to school, but it did yield results, and they were numeric, so people trusted them.

The argument for doing this was convincing – how could anyone argue against wanting to know where we stand on education? With this knowledge in hand, surely it would be easier to know where there were struggling schools and give them help.

But instead of coming to the aid of the school systems that needed help, we ended up punishing them, blaming the principals and teachers for the problems. The fact that the scores were extremely correlated to poverty was explained away by saying our teachers had “given up on poor students.”

Thus began the era of high stakes testing for students and teachers, where teachers were “held accountable” for their students’ progress on standardized tests, if not their scores. It’s purely punitive, and a far cry from the original purpose of helping out those who need it. We haven’t equalized funding for schools, for example.

Moreover, it hasn’t helped the students and schools which are struggling. In fact scores overall seem to be going up for everyone, but the rich students, to put it bluntly, are improving faster than the poor ones. And in the meantime everyone is getting sick of all the tests.

3. What’s next in educational reform?

Recent research has shown there might be a new explanation for why some kids do well when others don’t, and it’s all about “character,” or “grit.” Kids who have it in abundance seem resilient in the face of failure, and they thrive even in tough situations.

Hallelujah! Now we can try to develop kids’ grit quotients, or better yet, we can hold teachers accountable for doing so. The only problem with this plan is that it’s actually kind of hard to consistently measure grit.

Until now. Angela Duckworth, who is one of the social scientists that has been studying grit in children and who recently wrote an opinion piece on the subject, has helped develop a scorecard, based on self-assessment, so that kids can be scored.

In her piece, Duckworth warns us that the scoring system isn’t perfect, and she has even said it’s a bad idea to use this scoring system for accountability; it could undermine the very thing we’re trying to promote.

But she’s also said that kids can improve their scores, with help from teachers, and that high scores are good signs for progress in behavioral and academic achievement.

Let’s not go there.

I haven’t been blogging much lately, partly because I’ve been recovering from a stupid bike accident from last Friday, where I fell forward over my handlebars on the West Side bikepath, and partly because, as a result of my accident, I decided to stop and smell the roses a bit, meaning I actually read a novel. It was an amazing experience, reading a novel for the first time in years, especially one this distracting. It’s called Station Eleven, by Emily St. John Mandel, and if you have time, please read it.

I’m mostly recovered now, and I have to say, my one week of detox from the normal news cycle has been freaking amazing. I feel restored. Restored enough to read the recent Flint Water Advisory Task Force Report.

And, holy crap. It’s really good, and places the blame squarely on the State of Michigan, Governor Snyder, and in particular on the Emergency Management system that cares only about money over public health. I’ve blogged about the fucked up and racist system of Emergency Managers in Michigan before.

For a summary of the time-line of events which led to the widespread lead poisoning, take a look at pages 16 through 21. After that, if you want to get something else done this morning, jump to the following excerpt from page 54 which gets to the very heart of the issue:

Environmental justice embraces two fundamental principles: (1) the fair, non-discriminatory treatment of all people; and (2) the provision for meaningful public involvement of all people— regardless of race, color, national origin or income—in government decision-making regarding environmental laws, regulations and polices. Environmental justice or injustice, therefore, is not about intent. Rather, it is about process and results—fair treatment, equal protection, and meaningful participation in neutral forums that honor human dignity.

Environmental injustice is not about malevolent intent or deliberate attacks on specific populations, nor does it come in measures that overtly violate civil rights. Environmental injustices as often occur when parties charged with the responsibility to protect public health fail to do so in the context of environmental considerations.

The facts of the Flint water crisis lead us to the inescapable conclusion that this is a case of environmental injustice. Flint residents, who are majority Black or African American and among the most impoverished of any metropolitan area in the United States, did not enjoy the same degree of protection from environmental and health hazards as that provided to other communities. Moreover, by virtue of their being subject to emergency management, Flint residents were not provided equal access to, and meaningful involvement in, the government decision-making process.

The occurrence of environmental injustice in the Flint water crisis does not indict or diminish other public and private efforts to address Flint’s many challenging circumstances. However, irrespective of the intent of the parties involved, the simple reality is that the Flint water crisis is a case of environmental injustice.

Also, there’s this on page 56:

Among African American seniors, the protracted Flint water crisis echoes the tragic Tuskegee syphilis study and the decision not to treat smallpox among freedmen in the aftermath of the American Civil War. From this perspective, it is noted that measuring blood lead levels without removing the sources of lead from the environment—in this case, lead-tainted water—appears the equivalent of using Flint’s children (and adults) as human bioassays.

I’m so glad this work was done.