Qiming will be presenting this at ICML. If you're there, find him and say hi!

JHU Computer Science@jhucompsci.bsky.social · last mo.

In “Ski Rental with Distributional Predictions of Unknown Quality,” Qiming Cui and @mdinitz.bsky.social revisit the central online problem of ski rental ⛷️ in the “algorithms with predictions” framework from the point of view of distributional predictions: arxiv.org/abs/2602.21104 (2/25)

5-year postdoctoral research fellowship in theoretical Computer Science (also 4 other research fields), a nice opportunity! Needs PhD to be awarded between 1/Aug/24 and 1/Oct/27. Please share with whoever may be suitable! www.asc.ox.ac.uk/post-doctora...

this is a deep cut.

Standplaats Kraków could really use some cash!@standplaatskrk.pl · 2mo ago

Congratulations to Tal Rabin, Shubhangi Saraf, and Lisa Zhang, recognized by the SIGACT Distinguished Service Award for their role in making Theoretical Computer Science more welcoming and inclusive with the Women In Theory (WIT) workshop! sigact.org/prizes/servi... (via @gautamkamath.com)

Thanks to Sampath Rapuri, Alex Ma, and Edmund Sumpena for presenting their senior theses for us—we’re excited to see where your research will take you next!

The ACM Transactions on Algos now has the Harold N. Gabow Annual Best Paper Award, for research contributions with lasting significance in algos. Hal Gabow, an amazing algos researcher, was founding editor of TALG, and helped establish this award. Thanks Hal! (Submit yr best papers to TALG!) 1/n

Congrats to Irit Dveer Dinur of the Institute for Advanced Study & Weizmann Insitute, #NASmember Subhash Khot of New York University, Guy Kindler of @hebrewuniversity.bsky.social, Dor Minzer of @mit.edu, and Muli Safra of Tel Aviv University, winners of the 2026 Michael and Sheila Held Prize! (1/2)

I’m giving a talk at Stanford Theory Lunch this Thursday on recent work arxiv.org/abs/2511.16094 (joint work with awesome coauthors @mdinitz.bsky.social and Jeremy Fineman). Come by if you are around!

Learning-Augmented Online Algorithms for Nonclairvoyant Joint Replenishment Problem with Deadlines

This paper considers using predictions in the context of the online Joint Replenishment Problem with Deadlines (JRP-D). Prior work includes asymptotically optimal competitive ratios of $O(1)$ for the ...

arxiv.org

Postdoc positions available at the DSAI Institute at Johns Hopkins: ai.jhu.edu/careers/post... . If you're interested in either differential privacy or algorithms with predictions, please get in touch!

Wikipedia should have a bot that, when the time comes, automatically changes "[subject] is" to "[subject] was" etc. This bot should be known as the killer app.

Lately, I have been obsessed with developing theoretically based optimization algorithms that actually attain the best practical performance. Alas, the classic model of minimax optimal methods is overly conservative; it overfits to tune its worst-case. We found a path forward 1/

this is a pretty accurate summary of my research area

For any of my @princeton.edu friends and followers: there's a new policy out that seems like it would hurt the eating clubs, including my beloved Quad. Please consider registering your opposition! www.dontharvardourprinceton.com

Don't Harvard Our Princeton | Save Eating Clubs and Co-ops

A campaign opposing Princeton's new dining policy that could damage Princeton forever.

dontharvardourprinceton.com

And if you're faculty: I beg you, for the sake of everyone else in the world, strongly consider creating your own public webpages for your course instead of having everything locked up in Canvas.

First day of class for the semester. Told my 2.5-year-old that "Daddy is going to be a teacher today, just like your teachers!". Response: "Daddy you're so silly". Apparently I lack the gravitas of a daycare teacher.

Incredibly well deserved!!

Clément Canonne@ccanonne.github.io · last yr.

Huge congratulations to Tracy Kimbrel, who received the 2025 ACM SIGACT Service Award 🏆 for his time, dedication, and advocacy as Program Director for the Algorithmic Foundations (AF) program at the NSF! sigact.org/prizes/servi... #TCSSky

Synthetic Data is all the rage in LLM training, but why does it work? In arxiv.org/abs/2502.08924 we show how to analyze this question through the lens of boosting. Unlike boosting, however, our assumptions on the data and the learning method are inverted.

Escaping Collapse: The Strength of Weak Data for Large Language Model Training

Synthetically-generated data plays an increasingly larger role in training large language models. However, while synthetic data has been found to be useful, studies have also shown that without proper...

arxiv.org

A proof is a logical argument written to convince a skeptical audience. A corollary is that the best way to read a proof is to roleplay as a skeptical audience.

Interviewer: can you explain this gap in your resume? Networking researcher: UDP

Thomas House@tah-sci.com · 2y ago

Interviewer: Can you explain this gap in your resume? Epidemic modeller: Stochastic fade out.

Just saw a new result on the arxiv: polylog approximation for directed steiner tree! arxiv.org/abs/2412.10744 . Amazing breakthrough in a fundamental approximation algorithms problem. Congrats Bundit!

Breaking the Barrier: A Polynomial-Time Polylogarithmic Approximation for Directed Steiner Tree

The Directed Steiner Tree (DST) problem is defined on a directed graph $G=(V,E)$, where we are given a designated root vertex $r$ and a set of $k$ terminals $K \subseteq V \setminus {r}$. The goal is ...

arxiv.org

A statistic I've seen repeated in the discussion of health insurance is that UHC used an "AI with a 90% error rate" to deny claims. Recently I saw it repeated twice here: dailynous.com/2024/12/15/c... in which a moral philosopher struggles to figure out if murdering health insurance CEOs is moral.

Complications: The Ethics of the Killing of a Health Insurance CEO (guest post) - Daily Nous

The killing of UnitedHealthcare CEO Brian Thompson earlier this month has ignited moral debate around the world. Many have condemned the killing as they would any other murder. Others, though, have th...

dailynous.com

Prompt: Draw a picture of an American family at Thanksgiving dinner.

Just had a new paper hit the arxiv (will appear at NeurIPS '24), joint with Sungjin Im, Thomas Lavastida, Ben Moseley, Aidin Niaparast, and @vsergei.bsky.social: arxiv.org/abs/2411.16030 . I think it's super cool, so a quick thread!

Binary Search with Distributional Predictions

Algorithms with (machine-learned) predictions is a powerful framework for combining traditional worst-case algorithms with modern machine learning. However, the vast majority of work in this space ass...

arxiv.org

The Washington Post reports that Michael Anton and Sebastian Gorka had a falling out over who should receive credit for writing Trump's speeches. This is incorrect. In this thread, I will tell you the real reason why Anton dislikes Gorka. 🧵

Almost got through a semester with no academic integrity violations 😢