A Computer Scientist Who Tackles Inequality Through Algorithms

April 2, 2021 1:53 AM

Rediet Abebe uses the tools of theoretical computer science to understand pressing social problems — and try to fix them. Today, Abebe uses the tools of theoretical computer science to help design algorithms and artificial intelligence systems that address real-world problems. She has modeled the role played by income shocks, like losing a job or government benefits, in leading people into poverty, and she’s looked at ways of optimizing the allocation of government financial assistance. She’s also working with the Ethiopian government to better account for the needs of a diverse population by improving the algorithm the country uses to match high school students with colleges.

Abebe is a co-founder of the organizations Black in AI — a community of Black researchers working in artificial intelligence — and Mechanism Design for Social Good, which brings together researchers from different disciplines to address social problems.

posted by Alex404 (7 comments total)

Though it might seem off topic, this really chimes in an interesting way with some of the observations made by Nigerian-Irish UK academic Emma Dabiri in a recent conversation with Blindboy: the work of actually altering the structures that end up defining our societies, is exactly that phase that we are at risk of not seeing, given the high performativity component of recent social protests. The idea of engineering this coalition-building approach to addressing poverty by deconstructing it, so to speak, is so rare to get a glimpse of. The nuts and bolts of building a different, better, possible world. Thanks for the link.
posted by progosk at 4:11 AM on April 2, 2021 [1 favorite]

I kinda wish the quanta interviewer explained the lede in more detail, like how is theory used in addition to general CS research which already uses algorithms, optimization, proofs. Like what the difference is, especially since outside the field CS is perceived to be programming school which it is not.
posted by polymodus at 5:36 AM on April 2, 2021 [2 favorites]

Mechanism design is like if you had an algorithm designed, but you were aware that the input data is something that could be strategically manipulated. So you’re trying to create something that’s robust to that.

I'm so utterly glad that this is actually a field people work on. There clearly needs to be a lot more application of their work in the real world, but just reading this sentence made me relieved that this is a thing.
posted by automatronic at 7:42 AM on April 2, 2021 [1 favorite]

It is not just the data. The model matters.
posted by lalochezia at 8:39 AM on April 2, 2021

I kinda wish the quanta interviewer explained the lede in more detail, like how is theory used in addition to general CS research which already uses algorithms, optimization, proofs.

Well, the target audience of Quanta is people already familiar with CS research, so that is likely a given the author is free to assume. If CS folks want the general public to understand what they do, probably they can write their own articles and submit to PopSci?
posted by pwnguin at 10:04 AM on April 2, 2021 [1 favorite]

I hear the appetite as being for a bit more detail in the comp sci theory for that comp-sci-aware audience. Like is there maybe some causal logic or model fairness theorems or I don't know what going on, because that would be fascinating to understand more.

From the article the work sounds like it could be some of econometrics, ML fairness, algorithmic justice. Which are excellent, don't get me wrong, but if there's more theoretical theory, let's hear!
posted by away for regrooving at 5:28 PM on April 2, 2021

I kinda wish the quanta interviewer explained the lede in more detail, like how is theory used in addition to general CS research which already uses algorithms, optimization, proofs.

It's a fair point, and the Quanta article is indeed a bit on the thin side. You can get a little closer to some of the research questions by looking at the list of working groups on the Mechanism Design for Social Good website.

That being said, you'll notice on Rediet Abede's google scholar page that there's a pretty clear division between comp sci theory papers and... what's the word... humanitarian papers? I'd presume that the problem is partially that it's hard to get a theory paper published by motivating it with social goals. And the Quanta article didn't try to tease out some "hot new research" because the connection between them, although important, is still too tenuous.
posted by Alex404 at 11:07 PM on April 2, 2021 [1 favorite]