Dr. Jeff Jacobs

Who had the greatest influence on your education and/or career path?

Without a doubt, that would be my PhD advisor, Suresh Naidu!

From about age 12, until I started my PhD, I had walked around with this huge chaotic blob of cognitive dissonance in my head. On the one hand, I knew that learning how math and computers worked gave me nice little dollops of serotonin: my favorite thing in the world was (and is!) seeing some absolutely mind-blowing theorem or seemingly-magic-based algorithm and diving into it until I could finally capital-U Understand it to some extent. At the same time, growing up in the densely-packed but hyper-segregated DC area as a white male meant watching from one side of 16th St while my non-white friends on the other side faced stressors I couldn’t even imagine having to deal with: getting harassed by police, worrying about whether they’d have food to eat that night, clothes to wear to school the next day, and so on. So I knew I had immense privilege, and I knew that this privilege implied a corresponding responsibility to push back against these inequities.

Suresh provided me with a model of how one could be a responsible, “engaged” scholar. Reading his work and talking with him I think brought me towards a slow realization that I had completely goofed in seeing this as a zero-sum competition: an hour spent studying came at the cost of an hour I could be engaging with the real world, for example, and the weight of that tension would defuse my motivation to do either one. But, through the projects I worked on as a Ph.D. student within his lab, I slowly but surely began to see how these pursuits were mutually beneficial, erasing the all-or-nothing distinction between learning things as abstract theories and putting them into concrete practice. I immediately saw the relevance of how the data-scientific techniques I was learning could be used to study things I cared about in the world.

Suresh showed me, through his day-to-day life and his conduct as an advisor, scholar, and activist throughout all of these incredible projects, how it’s possible to be ethically and politically engaged without compromising scientific rigor and scholarly integrity. He taught me “The better you understand the world, the more effective you can be at changing it.”

What is computational political theory, and what drew you to study it?

The phrase “Computational Political Theory” came about in the same way that I’d guess new names for animals came about: I was trying to describe a type of bird I loved, and the person I was talking to would say “Oh, you mean an oriole?” And I’d say, no, not exactly, and they’d say “Oh, wait, you’re describing a raven!” And I’d say, no, that’s still not it, you’re just naming local sports team mascots now. (For my linguistics nerd friends: this is Quine’s “Gavagai” problem rearing its ugly head) Eventually—as my final paper for an incredible seminar class taught by Joshua Simon called Interpretation and Critique of Political Ideas—I started using that phrase to try and describe the theoretical part of my dissertation, where I was trying to flesh out what exactly it would mean for a computer to “read” a book or “examine” an archive.

Concretely, for example, I fell into a that-is-so-cool-I-must-study-it reverie during a class I took on the French Revolution, where the professor mentioned in passing something called the cahiers de doléances or ”grievance books”. These were essentially the world’s first-ever mass public opinion surveys, conducted in 1789 by King Louis XVI to try and figure out why everyone was so mad at him at the time. Everyone in the entirety of France—peasants, nobles, members of the clergy from every region in the country—gathered in local assemblies to deliberate and write down all of their grievances against the monarchy. The issue is, that this resulted in about 40,000 big books filled with grievances that nobody had the time to read in full, especially not in time to stem the tide of revolution. So, in place of thorough reading, politicians from then onwards “cherry-picked” through them, scooping out a biased selection of whatever grievances served their polemical purposes at the time, regardless of how representative they were of the concerns of the French populace more broadly.

So, the notion of using computers to conduct a statistically-representative study of these grievances instead is where, to me, this project (which became my thesis for the Quantitative Methods minor at Columbia) becomes even more exciting, since we can look at it as a fairly spot-on metaphor relating to modern-day issues like political polarization, media bias, misinformation, fake news, etc.: if we can come up with a way to have computers read and understand the grievances of various cross-sections of French society in 1789, incorporating and balancing all of the information without any pre-existing biases towards particular political viewpoints, perhaps that same kind of algorithm could be used to help us process the vast mountain of opinions which now get blasted in our face every day on social media, in newspapers, on TV, and so on! That’s just one example of what excites me about Computational Political Theory and, if it sounds at all interesting to anyone, they can check out this syllabus from a whole course on the subject that (sadly) ended up canceled a few weeks before it was set to start, due to the COVID pandemic.

What do you enjoy most about your work?

I genuinely don’t know where to start—in all honesty, it is my absolute dream job. But, one of the most rewarding aspects of teaching DSAN 5000 and 5100 specifically has been the opportunity to work with students on their final projects during office hours. Whether they’re describing an obstacle they’ve run into, or just hoping to brainstorm ideas for the next steps, I find myself getting completely engrossed in the conversations: along the way, I get to learn what motivates them, what their backgrounds are, what aspects of data science interest them in particular, what drew them to the particular dataset(s) they’re analyzing, and so on. It feels a bit like I’m at a party and everyone is telling me all about these cool new things I didn’t know about, but whereas at a party I have to write notes to myself for later (“google this when you get home!”), During office hours we get to dive directly into those things, exploring new datasets or algorithms and seeing how they can be incorporated into their projects.

What’s the first piece of advice you would give to a DSAN student?

That data science—and the math, computer science, and statistics that underlie it—doesn’t have to take the form of a bunch of symbols and equations that you stare at for a long time and then regurgitate onto a page when you take a test! That the scientific method gives us a general framework for understanding and critically evaluating the great mass of information we encounter on a day-to-day basis, and that probability and statistics in particular give us tools for integrating this new information with our previous experiences, allowing us to update our knowledge about things in the world in a systematic way (*cough* Bayes’ Rule *cough*).

What would people be surprised to learn about you?

As I get better at always being my full, authentic self around people I hope there are fewer and fewer surprises! But, given that I’m very far from that ideal, I feel like there may be a long list of things. Like…

I follow the NFL and NBA religiously and grew up in a household of rabid Washington Commanders fans and season ticket holders
I have an accidental ability to memorize song lyrics quickly so that I can sing along to a bunch of Regina Spektor or Imogen Heap or Sara Bareilles or Carly Rae Jepsen albums
My legal middle name is Power, which was less weird in its original form as it was my Irish great-grandfather’s last name

What’s the best advice you’ve ever received?

“Every topic in math is best learned twice, and even better learned three times!”

My undergrad advisor at the University of Maryland, William Gasarch, is a full-on superhero in my eyes. As a tenured professor, he took (and continues to take!) whatever free time he had and used it to help weirdo high school students like myself see how cool math is, no matter how terrible our grades were or how “bad” we thought we were at math. He is the type of person who can come up with sayings like this on the spot, right when they are most desperately needed to cheer people up. They were his first words to me after I came into his office dejected, tears welling up in my eyes, after failing MATH 404: Galois Theory (the course number is forever etched in my brain) for a second semester in a row.

If you could have one superpower, what would it be and why?

It’s going to come across like I’m dodging the question, but I feel like neural networks and other machine learning algorithms already give us sci-fi-level superpowers, the ability to do things far beyond what my younger computer/math nerd self could have ever imagined. Of the many examples that have cemented this view in my head, a big one would be the Artistic Style Transfer algorithms which started coming out in 2015: for decades, human creativity was held out as the one thing that computers could never replace, at the core of what it meant for us to be human, and suddenly one day someone posts a paper showing how you can input a photo, tell the algorithm “paint this like Van Gogh… no, wait, like Seurat!”, and have their goofy selfie or cute photo of their pet instantly rendered as a beautiful work of art.

What 3 things would you want with you on a deserted island?

My French Bulldog Kozo, my cat Biko, and… probably the brain of a Zoomer, since getting stranded on a desert island is extremely fun to Zoomers: they pay money to spend hours and hours doing it, every day, and it’s called Minecraft.