Thiago Serra

Those irrelevant perturbations are smaller but still present in tabular foundation models, as observed with in-context learning over TabPFN.

Yang suggests this as a natural limitation on the use of this type of model for this type of task, which is a reasonable perspective.

4/4

To investigate if such variations could be due to randomness instead of lack of robustness, at least in the simpler but less competitive case of in-context learning, Yang evaluated attention scores in the open model Llama-38-B. The scores varied by row order, variable names, & variable values.

3/4

In the simple setting of a linear function as ground truth, Yang found that supervised fine-tuning of gpt-4o is competitive with commonly used regression techniques—although being an absolute overkill. However, it is sensitive to task-irrelevant variations, such as change in number of digits.

2/4

What can’t LLMs do? If you get caught using them for data fitting, you deserve a warning.

Mochen Yang talked at Iowa’s Tippie College of Business about the caveats of using LLMs for that purpose. After all, getting gpt-4o to predict 0.1875 entails predicting tokens “0”, “.”, “187”, and “5”!

1/4

Editorial Board | INFORMS Journal on Computing Editor-in-Chief: Alice E. Smith, Auburn University; Area Editors listing; Assoc. Editors listing; Advisory Board; Former Editors-in-Chief listing

As someone who has always been involved with computing, optimization, and AI…

I am happy to share that I am joining the editorial board of the INFORMS Journal on Computing as Associate Editor in AI and Optimization! 😎

pubsonline.informs.org/page/ijoc/ed...

Springer gives money for best paper awards - at least to CPAIOR

Summer 2026 schools on artificial intelligence, data science, machine learning, and operations research The purpose of organizing this list is to help graduate students find a summer school to gain skills related to operations research. Why? Because students are often unaware of those, whereas facult…

The Summer 2026 list of schools has been updated with four new opportunities. You can look for those directly by searching for "Included on January 15"!

thiagoserra.com/2026/01/04/s...

Title + abstract of the preprint

Excited to present a new preprint with @nkgarg.bsky.social: presenting usage statistics and observational findings from Paper Skygest in the first six months of deployment! 🎉📜

arxiv.org/abs/2601.04253

Summer 2026 schools on artificial intelligence, data science, machine learning, and operations research The purpose of organizing this list is to help graduate students find a summer school to gain skills related to operations research. Why? Because students are often unaware of those, whereas facult…

Kicking off 2026, here is a new list of summer schools on artificial intelligence, data science, machine learning, and operations research:

thiagoserra.com/2026/01/04/s...

The driving force behind this paper is Hung Pham, who kept pushing it for three semesters after working with me in my last summer at Bucknell University. He is currently applying to PhD programs in operations research and related disciplines. Watch out for Hung’s application!

Optimization over Trained (and Sparse) Neural Networks: A Surrogate within a Surrogate In constraint learning, we use a neural network as a surrogate for part of the constraints or of the objective function of an optimization model. However, the tractability of the resulting model is he...

I am happy to share a new version of the Surrogate Within A Surrogate paper. This paper shows that we can find better solutions for an optimization problem involving a dense neural network by pruning it. We made an extra effort to clarify our contributions and to expand our experimental evaluation.

Not that successful in tweaking this with AI…

Make a Bond movie academic:

Facets are Forever

Facets are Forever? 💎 📝

Make a Bond movie academic:

(Sean Connery is unfunded)

You Only Travel Twice

Make a Bond movie academic:

License to Queue

#orms

Make a Bond movie academic:

"The World is Not Enough: A PhD dissertation on the avarice of billionaires and the socio economic divide causing a fracture in democratic processes”

Looking for an auspicious PhD? Try it the Iowa way!

We have a top-notch department with active research at the intersection of many important fundamental and emerging subjects. There is still time—the application deadline is three weeks away! More information here: tippie.uiowa.edu/phd/phd-busi...

Happy to have contributed to Anna Mitchell’s cover story at The Daily Iowan about data centers in Iowa.

You can read it here: dailyiowan.com/2025/12/09/c...

When Deep Learning Meets Polyhedral Theory: A Survey In the past decade, deep learning became the prevalent methodology for predictive modeling thanks to the remarkable accuracy of deep neural networks in tasks such as computer vision and natural langua...

In related news, my coauthors and I recently got the news that our survey (my third paper in citations) will be soon published at the INFORMS Journal on Computing.

For now, you can check the preprint here: arxiv.org/abs/2305.00241

Today I am celebrating the symbolic thousandth citation on Google Scholar.

Next week, I will be participating on a panel hosted by @gurobioptimization.bsky.social about training the next generation of mathematical optimizers.

You can join the conversation here: www.brighttalk.com/webcast/1918...

I’m excited to share the materials from my Stanford seminar course, “AI for Algorithmic Reasoning and Optimization”: vitercik.github.io/ai4algs_25/. It covered formal algorithmic frameworks for analyzing LLM reasoning, GNNs for combinatorial/mathematical optimization, and theoretical guarantees.

It's been an absolute pleasure working with Ellen, Madeleine, and their amazing PhD students for the past year on making optimization more accessible with generative AI!

I am on the job market this year - check out my website (conlaw.github.io) for more details on what I've been up to.

Please keep an eye out for Connor Lawless (@lawlessopt.bsky.social) on the faculty job market! Connor is a Stanford Human-Centered AI Postdoc, co-hosted by myself and Madeleine Udell. His research combines ML, computational optimization, and HCI, with the goal of building human-centered AI systems.

The 2025 ICS Student Paper Award winner is Hao Hao (CMU) for the paper "Robust Paths: Geometry and Computation," co-authored with Peter Zhang.

The 2025 ICS Harvey J Greenberg Research Award was presented to Braden L. Crimmins, J. Alex Halderman, and Bradley Sturt for their paper "Improving the Security of United States Elections with Robust Optimization."

The 2025 ICS Prize Committee recommended the recognition of two submissions as co-winners of the ICS Prize:
(1) Alper Atamtürk, Andrés Gómez, and Shaoning Han
(2) Jason Altschuler and Pablo Parrilo

ok

And that’s a wrap for #informs2025

See you next year in San Francisco!