Natalie Collina

Why do I have to pretend that I'm going to print something in order to save it as a PDF. Why do I have to engage in a little ruse.

Heard joke once: Man goes to doctor. Says he's depressed. Says life seems harsh and cruel. Says he feels all alone in a threatening world where what lies ahead is vague and uncertain. Doctor says, "Sorry: you have exceeded your token usage. Please try again later or switch to Auto"

Tomorrow's front page of the Minnesota Star Tribune: Jan. 24, 2026

Excited about a new paper! Multicalibration turns out to be strictly harder than marginal calibration. We prove tight Omega(T^{2/3}) lower bounds for online multicalibration, separating it from online marginal calibration for which better rates were recently discovered.

Distraught woman says ICE killed her wife in video after deadly Minneapolis shooting "They killed my wife," the distraught woman says, adding, "They shot her in the head." An ICE agent shot and killed a 37-year-old woman who was driving an SUV in Minneapolis on Wednesday. / Screenshot/@Breaking911

"They killed my wife. I don't know what to do," the woman says through sobs in the footage, with a damaged SUV visible in the distance behind her. "We stopped to videotape, and they shot her in the head," the woman cries. "We have a six-year-old at school," she says, almost unable to breathe, as a chaotic scene in which federal officers prevented at least one doctor who was on the scene from assisting the shot victim unfolds. "We're new here," the distraught woman says in despair.

You and your wife drop your 6-year-old off at school. You just moved here. You see ICE terrorizing your new neighbors. You film them, as is your legal right. Your wife complies with orders. She is then shot in the head. You still have to pick up your child later today.

This could be you.

Image is a popular astronaut meme. An astronaut is looking at the earth from space saying wait, I am worthy of love? An astronaut is behind him holding up flowers and saying always have been

A message for you all to carry through the year

I just think it would be neat if turkeys used the same technique to draw us

TCS+ RSVP: Natalie Collina (2025/12/03) Title: Swap regret and correlated equilibria beyond normal-form games

📢 Our last TCS+ talk of the season will be Wed, Dec 3 (10am PT, 1pm ET, 19:00 CET): Natalie Collina (@ncollina.bsky.social), from UPenn, will tell us about "Swap regret and correlated equilibria beyond normal-form games"!

RSVP to receive the link (one day before the talk): forms.gle/utLgSxLpqvpx...

My 11-year-old sitting with her pile of Halloween candy, sorting it into a bar graph

We have progressed from data collection to data analysis.

Venn diagrams as jack-o-laterns showing trick/treat logicals: OR, AND, XOR, NOR, NAND, and XNOR

Halloween logicals, still the best 10/31 venn

English Oak

Nature’s Gold, England’s Glory

Always happy to chat about it if you wanna get in the game! 😊

Yeah it’s hard to regulate properties of algorithms, especially as they get increasingly complex! I agree abt market interventions. We have a followup work (should be up in a few days!) about how larger market sizes and a small number of entrants applying simple heuristics helps mitigate collusion.

It’s true!

And now might be a good time to mention, I’m on the faculty job market this year! I do work in human-AI collusion, collaboration and competition, with an eye towards building foundations for trustworthy AI. Check out more info on my website here! Nataliecollina.com

The Game Theory of How Algorithms Can Drive Up Prices | Quanta Magazine Recent findings reveal that even simple pricing algorithms can make things more expensive.

Our paper on algorithmic collusion was featured in a Quanta article! www.quantamagazine.org/the-game-the...

Appearing in SODA 2026! Last year we had a 3-page SODA paper, this one is 107 pages. Next time I’m thinking we swing way back the other way and just submit a twitter/bluesky thread

Aligning an AI with human preferences might be hard. But there is more than one AI out there, and users can choose which to use. Can we get the benefits of a fully aligned AI without solving the alignment problem? In a new paper we study a setting in which the answer is yes.

Thing with Grok is that better versions will be less overt and more convincing. An expert on arguing about the deficit, immigration, public health, etc. with a particular political slant

This is sort of obviously true for how Elon interacts with grok, but I think will be increasingly true more broadly. Next time you get a weird answer from an LLM, ask yourself, would the company that owns this tool like this answer?

As AI systems continue to become more complex and hard to reason about, one increasingly powerful lens through which to understand them is through the incentives of AI creators.

Ecstatic and deeply honored by this award. I've had great fun thinking about algorithms as strategies for repeated games over the past few years and hope that this highlight will push more researchers to come up with exciting directions in this field! Come to our talk on Monday to learn more!

This best paper news is a good opportunity to highlight that a month or so ago I started maintaining CV of failures on my website. It will almost certainly continue to grow linearly in the number of things I attempt to do, and that’s a good thing! www.seas.upenn.edu/~ncollina/Fa...

Anyway, if you made it all the way to the end of this thread, thank you so much for reading! Come learn more about the paper this Monday at EC! ✨

Specifically, in Bayesian games, there are payoff profiles you can induce via two non-manipulable algorithms playing against each other which are *impossible* to attain via a mediator providing correlated action recommendations that agents best-respond to!

These perspectives are the same in normal-form games, but it turns out they’re meaningfully DIFFERENT even if you move just to Bayesian games!

So in normal form games we can actually define CE in two equivalent ways: one is as any outcome induced by two non-manipulable algorithms playing against each other. One is as the outcome of a mediator model, where a third party sends correlated signals to each agent and each agent best-responds.

This also leads to a really cool result about Correlated Equilibria (CE)! In addition to having many strategic properties, swap regret has a tight connection to CE in normal form games; the set of move pair distributions that can be induced by two swap regret algorithms is exactly the set of all CE

By using cutting-edge tools from approachability, we are able to efficiently approach this non-manipulable set even though deciding membership is hard. And the algorithm that does so is an efficient no-Profile Swap Regret algorithm!

So what’s our trick? Instead of framing the problem as searching for the right “swap” function in these highly complex games, we directly work in menu space, where the property of non-manipulability has a very natural interpretation: the extreme points of the menu are product distributions!