Weapons of math destruction and the ethics of Big Data

I don’t know about you, dear reader, but during my formal education I was never taught ethics or social consciousness. I even remember sitting around with my engineering friends that had to take a class in ethics and laughing at the irrelevance and futility of it. To this day, I have a strained relationship with ethics as a branch of philosophy. However, despite this villainous background, I ended up spending a lot of time thinking about cooperation, empathy, and social justice. With time and experience, I started to climb out of the Dunning-Kruger hole and realize how little I understood about being a useful member of society.

One of the important lessons I’ve learnt is that models and algorithms are not neutral, and come with important ethical considerations that we as computer scientists, physics, and mathematicians are often ill-equipped to see. For exploring the consequences of this in the context of the ever-present ‘big data’, Cathy O’Neil’s blog and alter ego mathbabe has been extremely important. This morning I had the opportunity to meet Cathy for coffee near her secret lair on the edge of Lower Manhattan. From this writing lair, she is working on her new book Weapons of Math Destruction and “arguing that mathematical modeling has become a pervasive and destructive force in society—in finance, education, medicine, politics, and the workplace—and showing how current models exacerbate inequality and endanger democracy and how we might rein them in”.

I can’t wait to read it!

In case you are impatient like me, I wanted to use this post to share a selection of Cathy’s articles along with my brief summaries for your browsing enjoyment:

Models, modesty, and moral methodology

In highschool, I had the privilege to be part of a program that focused on the humanities and social sciences, critical thinking, and building research skills. The program’s crown was a semester of grade eleven (early 2005) dedicated to working on independent research for a project of our own design. For my project, I poured over papers and books at the University of Saskatchewan library, trying to come up with a semi-coherent thesis on post-cold war religious violence. Maybe this is why my first publications in college were on ethnocentrism? It’s a hard question to answer, but I doubt that the connection was that direct. As I was preparing to head to McGill, I had ambition to study political science and physics, but I was quickly disenchanted with the idea, and ended up focusing on theoretical computer science, physics, and math. When I returned to the social sciences in late 2008, it was with the arrogance typical of a physicist first entering a new field.

In the years since — along with continued modeling — I have tried to become more conscious of the types and limitations of models and their role in knowledge building and rhetoric. In particular, you might have noticed a recent trend of posts on the social sciences and various dangers of Scientism. These are part of an on-going discussions with Adam Elkus and reading the Dart-Throwing Chimp. Recently, Jay Ulfelder shared a fun quip on why skeptics make bad pundits:

First Rule of Punditry: I know everything; nothing is complicated.

First Rule of Skepticism: I know nothing; everything is complicated.

Which gets at an important issue common to many public-facing sciences, like climate, social, or medicine, among others. Academics are often encouraged to be skeptical, both of their work and others, and precise in the scope of their predictions. Although self-skepticism and precision is sometimes eroded away by the need to publish ‘high-impact’ results. I would argue that without factions, divisions, and debate, science would find progress — whatever that means — much more difficult. Academic rhetoric, however, is often incompatible with political rhetoric, since — as Jay Ulfelder points out — the latter relies much more on certainty, conviction, and the force with which you deliver your message. What should a policy oriented academic do?

Cross-validation in finance, psychology, and political science

A large chunk of machine learning (although not all of it) is concerned with predictive modeling, usually in the form of designing an algorithm that takes in some data set and returns an algorithm (or sometimes, a description of an algorithm) for making predictions based on future data. In terminology more friendly to the philosophy of science, we may say that we are defining a rule of induction that will tell us how to turn past observations into a hypothesis for making future predictions. Of course, Hume tells us that if we are completely skeptical then there is no justification for induction — in machine learning we usually know this as a no-free lunch theorem. However, we still use induction all the time, usually with some confidence because we assume that the world has regularities that we can extract. Unfortunately, this just shifts the problem since there are countless possible regularities and we have to identify ‘the right one’.

Thankfully, this restatement of the problem is more approachable if we assume that our data set did not conspire against us. That being said, every data-set, no matter how ‘typical’ has some idiosyncrasies, and if we tune in to these instead of ‘true’ regularity then we say we are over-fitting. Being aware of and circumventing over-fitting is usually one of the first lessons of an introductory machine learning course. The general technique we learn is cross-validation or out-of-sample validation. One round of cross-validation consists of randomly partitioning your data into a training and validating set then running our induction algorithm on the training data set to generate a hypothesis algorithm which we test on the validating set. A ‘good’ machine learning algorithm (or rule for induction) is one where the performance in-sample (on the training set) is about the same as out-of-sample (on the validating set), and both performances are better than chance. The technique is so foundational that the only reliable way to earn zero on a machine learning assignments is by not doing cross-validation of your predictive models. The technique is so ubiquotes in machine learning and statistics that the StackExchange dedicated to statistics is named CrossValidated. The technique is so…

You get the point.

If you are a regular reader, you can probably induce from past post to guess that my point is not to write an introductory lecture on cross validation. Instead, I wanted to highlight some cases in science and society when cross validation isn’t used, when it needn’t be used, and maybe even when it shouldn’t be used.

Big data, prediction, and scientism in the social sciences

Much of my undergrad was spent studying physics, and although I still think that a physics background is great for a theorists in any field, there are some downsides. For example, I used to make jokes like: “soft isn’t the opposite of hard sciences, easy is.” Thankfully, over the years I have started to slowly grow out of these condescending views. Of course, apart from amusing anecdotes, my past bigotry would be of little importance if it wasn’t shared by a surprising number of grown physicists. For example, Sabine Hossenfelder — an assistant professor of physics in Frankfurt — writes in a recent post:

It isn’t so surprising that social scientists themselves are unhappy because the boat of inadequate skills is sinking in the data sea and physics envy won’t keep it afloat. More interesting than the paddling social scientists is the public opposition to the idea that the behavior of social systems can be modeled, understood, and predicted.

As a blogger I understand that we can sometimes be overly bold and confrontational. As an informal medium, I have no fundamental problem with such strong statements or even straw-men if they are part of a productive discussion or critique. If there is no useful discussion, I would normally just make a small comment or ignore the post completely, but this time I decided to focus on Hossenfelder’s post because it highlights a common symptom of interdisciplinitis: an outsider thinking that they are addressing people’s critique — usually by restating an obvious and irrelevant argument — while completely missing the point. Also, her comments serve as a nice bow to tie together some thoughts that I’ve been wanting to write about recently.