Thiago Serra

17 Feb

I am happy to share about a new paper with Jiatai Tong, Yilin Zhu, and Sam Burer — which also just got accepted at CPAIOR 2026: Optimization over Trained Neural Networks: Going Large with Gradient-Based Algorithms 1/N

arxiv.orgOptimization for One, Feasibility for Most: Decoupling Optimization from Feasibility in Chance-Constrained ProgramsChance-constrained stochastic programs (CCSPs) require solutions that satisfy uncertain constraints with high probability. Standard approaches either embed scenarios directly into optimization by cons...

The preprint of "Optimization for One, Feasibility for Most: Decoupling Optimization from Feasibility in Chance-Constrained Programs", coauthored with Justin Dumouchelle, Shunyu Yao, and Beste Basciftci, is now on arXiv: arxiv.org/abs/2610.19027

Thiago Serra@thserra.bsky.social

Wrapping up #cors2026, Justin Dumouchelle talked about our joint work (along with Shunyu Yao and Beste Basciftci) on decoupling optimization from feasibility when solving chance-constrained optimization models. 1/N

Selvaprabu (Selva) Nadarajah is visiting the University of Iowa Tippie College of Business and giving a Business Analytics seminar on solving real options problems for energy options, which concerns trading on energy investments, using Markov Decision Process models.

View full thread

Nadarajah presented a stream of worked coauthored with Andre Augusto Cire, Qihang Lin, Negar Soheili, and Bo Yang on…

… rather adapting the fidelity of the approximate solution of the MDP to the computational resources available.

There is often a tradeoff between the fidelity of academic models vs. the tractability and interpretability of the models solved for making such decisions in practice.

Thiago Serra reposted

Paper Skygest Team

@paper-feed.bsky.social

19 Aug 2025

**Please repost** If you're enjoying Paper Skygest -- our personalized feed of academic content on Bluesky -- we'd appreciate you reposting this! We’ve found that the most effective way for us to reach new users and communities is through users sharing it with their network

mathandai.orgDeclaration — Math and AIRead the declaration and add your name.

12 Sept

I am endorsing this declaration about the misalignment between AI and mathematics: mathandai.org

12 Sept

“ In recent months, the success of AI in solving major mathematical problems has made headlines even outside mathematical circles. But solving problems is only a tool and proxy for achieving the primary goal of conceptual understanding and insight.

@angelamczhou.bsky.social

12 Sept

Forgetting this in the world of AI may turn the tool against the primary goal. Indeed, the mass production at faster and faster pace of "true/false" statements could destroy fertile ground instead of breathing life into new ideas. “

Thiago Serra reposted

angela zhou

11 Sept

✊

Terence Tao@teorth.bsky.social

A group of 25 Fields Medalists, including myself, have made a joint declaration on Math and AI: mathandai.org . We welcome additional signatories. See also this article in the Economist announcing the declaration: www.economist.com/science-and-...

mathandai.orgDeclaration — Math and AIRead the declaration and add your name.

25 Aug

One missing post from last week’s #FutureBAProf workshop: Violet (Xinying) Chen talking about surviving the tenure track. Chen talked first about understanding expectations, building a balanced publication portfolio, networking for visibility, and knowing when to change plans.

25 Aug

Later she also usual mistakes such as having too many projects, waiting too long to submit papers, and advise students alone only if needed and not treating them as research assistants.

19 Aug

Wrapping up #FutureBAProf, Gilbert Laporte and Ahmed Abbasi discussed the future of Business Analytics. They both emphasized the importance of students to read as much as they can to become better scholars.

Today we have a special session at the #FutureBAProf workshop in memory of Warren Boe, a long time member and benefactor of the Business Analytics Department at The University of Iowa Tippie College of Business.

Ann Melissa Campbell and Kang Zhao shared stories and appreciation for having Boe as a colleague.

Qi Luo did such an amazing job last year at #FutureBAProf workshop that he is back again to talk about the academic job market from the applicant perspective

Thiago Serra@thserra.bsky.social

Qi Luo presented the applicant perspective about the academic job market at the #FutureBAProf workshop with a personal perspective on having switched from a position in industrial engineering to another one in business analytics. 1/2

Wrapping up the first day of talks at #FutureBAProf, Qi Feng talked about the on campus interview: what schools are looking for, what proxies they use to assess the candidates, and how the candidates can demonstrate their potential

(A few more slides)

Qihang Lin and Violet (Xinying) Chen talked at #FutureBAProf about academic job applications and conference interviews.

They stressed the importance of precision in the CV (status of each paper; authorship order), how to structure other documents, and how to prepare an elevator pitch to encourage follow up from interviewers.

Gilbert Laporte talked about how to structure a paper at #FutureBAProf Among many pearls and some jokes, the advice on how to write the introduction was very interesting: it should answer the “what?” and “why?”

Continuing the Research in Business Schools session at #FutureBAProf, Ahmed Abbasi talks about the fast-pace of change in scholarship on human-centered computational design with the emergence of inferential, generative, and agentic models.

Abbasi stressed the importance of abstraction in scholarship, so that research projects are not centered on temporal issues that will get quickly saturated by the first ones to get a paper published on that very specific topic.

He also pointed out at a trend of papers relying more on existing datasets rather than collecting their own. As factors of standing out as a scholar (and in the job market), he stressed abstraction, rigor, and domain depth with proper data collection.

Starting the Research in Business Schools session at #FutureBAProf, Annabelle Feng talks about the importance of thinking of your long-term research agenda: trends come and go, along with editorial priorities of journals, but the important questions remain, and they center on how we make decisions.

Feng emphasized how the way we approach problems change over time, that you should learn about new methods even if you are not currently planning on using them, and how there is always a model underneath even if our focus is on the data.

Feng talked about decision-focused learning as one of the current streams of the discipline, and about how establishing a uniform validity layer for how different methods are used is what will establish the maturity of the business analytics discipline.

We are kicking off the 5th edition of the #FutureBAProf workshop with Ann Melissa Campbell’s opening and Sam Burer’s introduction to business analytics.

This year we have brought 22 students from 21 institutions out of 117 applicants, including PhD students in second half of their PhDs as well as postdocs from business schools, CS, and engineering departments.

They will learn about the academic career in business analytics departments, how to navigate the job market, and how to thrive in their tenure track years from Ann Melissa Campbell, Sam Burer, Qi Feng, Ahmed Abbasi, Gilbert Laporte, Qihang Lin, Beste Basciftci, Tallys Yunes, Qi Luo, and Violet Chen.

youtu.beOptimal Combinatorial Testing with Constraints: The Balancing ActYouTube video by Thiago Serra

22 July

I am happy to share a new paper on optimization for combinatorial testing with Changkun Guan, Hunter Gehman, Sumit Dhar, Mikey Ferguson, John Hooker, and Marcel Schoppers. Link in the comments. This is one of the most fun optimization problems that I have worked on!

arxiv.orgOptimal Combinatorial Testing with Constraints: The Balancing ActImagine that you are in front of a cockpit with several on-off buttons. If you were to thoroughly test it, you would need to try a prohibitive number of configurations. But since most bugs in practice...

22 July

Link to arXiv version: arxiv.org/abs/2607.17083

optimization-online.orgOptimal Combinatorial Testing with Constraints: The Balancing Act – Optimization Online

22 July

Link to OO version: optimization-online.org/2026/07/opti...

25 June

I wrote a piece for ORMS Tomorrow called "PhD: My mistakes, my lucky breaks, and what I tell my students now". This article draws from my own experience as a student & from what I have learned by advising PhD students so far. Thank you Dan Ni Lin for the invitation! www.informs.org/content/down...

Thiago Serra reposted

Sophie Huiberts

@sophie.huiberts.me

25 June

MIP Europe 2026 will feature a poster session. If you wish to participate, make sure to register by July 10th. www.mixedinteger.org/EUROMIP/2026... (Poster submission deadline itself is August 31, July 10 is deadline to register)

20 June

But if I were to stay here for longer than 4 hours, would Google Scholar notice it?

As a result, we explore solutions at the feasibility boundary, the results are consistently above the intended feasibility rate, and the runtime is considerably smaller than the standard approach. 6/N

12 June

Our preprint will be out soon! 7/7

We address that as follows: A) We solve the optimization problem using a single scenario. B) To avoid overfitting, we multiply the right-hand side of the constraint by a parameter. C) We adjust the value of this parameter by cheaply using a massive set of scenarios to evaluate feasibility. 4/N

12 June

D) We use cheaper alternatives, including an heuristic and a graph neural network, to make the feasibility check faster. 5/N

Chance-constrained optimization has peculiar challenges: 1) If we optimize while ensuring feasibility, such as with sample average approximation, the feasibility gets slightly overfitted to the scenarios used. 2/N

12 June

2) In the case of binary variables, most solutions satisfy either all or none of the constraints. Hence, training a machine learning model to predict feasibility requires careful curation of its training set. 3/N

10 June

Carlos Zetina talked at #cors2026 about the benefits of GPU acceleration in solving linear programs using the FICO Xpress solver. Their solver has the two most popular methods implemented: - PDHG is faster in the problems that it can solve - cuPDLP+ solves more problems 1/2

10 June

He share interesting observations: - only delegating matrix multiplication to the GPU doesn’t yield a significant speedup - gaming GPUs are enough to obtain good results - an LP with more than 10M coefficients is solved faster in GPUs - there is a lot yet to be done for solving MILPs with GPUs 2/2

One use case of machine learning in optimization is circumventing the cost of optimization algorithms. Jerry Sun presented such a case at #cors2026 by predicting Farkas multipliers to prove that a solution is infeasible with respect to the linear relaxation without solving the linear relaxation.

Seyedeh Parisa Moosavi talked at #cors2026 about relaxations for weakly-coupled Markov decision processes. She considered an example of managing fatigue in crew scheduling: teams need a break to recharge, but their breaks need to be coordinated with the breaks of other teams. 1/2

Existing approaches, such as Lagrangian relaxation and approximate dynamic programming, end up representing unreachable states. By augmenting the arcs labels with the history of nodes previously visited, she manages to improve performance with little impact on computational costs. 2/2