LightNews
thserra.bsky.social
Thiago Serra
@thserra.bsky.social
3.1K followers 1.1K following 590 posts
Assistant professor at University of Iowa, formerly at Bucknell University, mathematical optimizer with an #orms PhD from Carnegie Mellon University, curious about scaling up constraint learning, proud father of two
Posts Media Videos Starter Packs
Pinned
thserra.bsky.social
Thiago Serra @thserra.bsky.social · May 6
We have a new paper out, with some exciting results on how to solve optimization models embedding neural networks!

📝 Optimization over Trained (and Sparse) Neural Networks: A Surrogate within a Surrogate

🔗 arxiv.org/abs/2505.01985

🎤 Talk in Montreal tomorrow: www.crmath.ca/en/activitie...

1/N
Optimization over Trained (and Sparse) Neural Networks: A Surrogate within a Surrogate
We can approximate a constraint or an objective function that is uncertain or nonlinear with a neural network that we embed in the optimization model. This approach, which is known as constraint learn...
arxiv.org
1 11
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 5h
Iowa City folks:

Tomorrow(Friday, October 17) I will be talking at the University of Iowa Computer Science's Colloquium @uiowacs.bsky.social about optimization over trained neural networks.

The talk is at 3:30 PM in room 110 MLH.

For more details: cs.uiowa.edu/event/34670/0
2
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 1d
www.smbc-comics.com/comic/apply
Saturday Morning Breakfast Cereal - Apply
Saturday Morning Breakfast Cereal - Apply
www.smbc-comics.com
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 1d
smbc-comics.com/index.php?id...
Saturday Morning Breakfast Cereal - Ur
Saturday Morning Breakfast Cereal - Ur
smbc-comics.com
1 2
Reposted by Thiago Serra
primaldual.bsky.social
Erling D. Andersen @primaldual.bsky.social · 1d
@sophie.huiberts.me is featured in a Quanta magazine article about the simplex method.

www.quantamagazine.org/researchers-...
Researchers Discover the Optimal Way To Optimize | Quanta Magazine
The leading approach to the simplex method, a widely used technique for balancing complex logistical constraints, can’t get any better.
www.quantamagazine.org
5 15
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 2d
A little tease:
🔍 Resume Matching – Are standardized resumes advantaged by AI tools?
📂 Resume Classification – How do we stop past biases from repeating?
🤖 Large Language Models – Can summaries distort or reward prompt tricks?
📜 Humanities Perspective – Do hiring tools embed cultural values & power?
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 2d
AI is reshaping hiring—but how exactly?

In this IJHRM editorial with Melissa Intindola, Neil Boyd, and John Hunter, we explore why understanding AI’s impact on hiring requires both technical insight and HR expertise.

www.tandfonline.com/doi/full/10....
Interdisciplinary narratives on artificial intelligence & personnel selection systems
Published in The International Journal of Human Resource Management (Vol. 36, No. 14, 2025)
www.tandfonline.com
1 2
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 3d
@jannisku.bsky.social is giving a very interesting talk on explainable (integer) optimization at the @euroorml.bsky.social seminar, highlighting counterfactual explanations before diving on new work with Coherent Local Explanations for Mathematical Optimization (CLEMO): arxiv.org/abs/2502.04840
3 8
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 3d
It was great having Joe Paat visiting The University of Iowa Tippie College of Business on Friday — and having a chance to further understand this work!
6
Reposted by Thiago Serra
nextmv.bsky.social
Nextmv @nextmv.bsky.social · 7d
What do @ryanjoneil.bsky.social and @thserra.bsky.social think is worth knowing about translating operations research and decision science learnings into industry practice? Join this session to find out: www.nextmv.io/videos/what-...

#orms #informs #operationsresearch #decisionscience #datascience
What to know about practicing operations research and decision science in industry
Two operations research PhDs with varied real-world experiences explore skillsets, actions, and considerations for entering into and practicing in OR and decision science in industry analytics setting...
www.nextmv.io
2 3
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 9d
🔥🔥🔥
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 10d
Last, but not least, I also managed to see how fast my little niece is growing during my quick hop connecting in Sao Paulo! 😍

3/3
2
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 10d
This was a great opportunity for strengthening my ties with the Brazilian AI community and meet many talented colleagues. I am very grateful for the invitation by Rosiane de Freitas-Rodrigues to play such an important role at the conference, and for reconnecting me with my academic origins.

2/3
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 10d
Last week I had the immense privilege of giving the opening keynote talk of the BRACIS - Brazilian Conference on Intelligent Systems #bracis2025 and visiting the beautiful city of Fortaleza for the first time.

1/3
1 2
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 12d
Second, she has partnered with SAS Brasil to first train people who will produce a better and more inclusive melanoma dataset. With a sensitization of the human who diagnoses about what types of lesions are malign on different skin tones, she hopes that we will eventually train the right model!

5/N
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 12d
The recognition of her work led to new and unexpected references and collaborations. First, the cosmetic industry is considerably ahead in cataloging skin tones, and in fact celebrated her contributions.

4/N
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 12d
On the other hand, those models missed the mark in lesions at body extremities like the hand palm, foot sole, and under the nails; which have considerably higher incidence in darker-skinned individuals.

3/N
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 12d
By removing the skin lesions entirely from the images in one way or another, her team still obtained models that were still considerably accurate due to other characteristics present in those images.

2/N
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 12d
Closing #bracis2025 with mastery, Sandra Avila talked about her journey working on diagnosing melanomas (a type of skin cancer).

Interestingly, the story starts with winning a classification contest, and then evolves from her working backwards to figure out what the models actually predicted.

1/N
1 2
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 15d
Although such attacks can be successful without interfering directly with the node being attacked, they lead to high-rank tensors. Hence, low-rank tensor approximations end up being a way to filtering the attacker’s noise from the information encoded in the graph.

6/6
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 15d
Towards the end, Papalexakis talked about computationally efficient clustering methods, and how they may be bound to adversarial attacks in graph neural networks.

5/N
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 15d
Rank-one tensors are blocks obtained as the product of vectors, which approximate part of the information in the tensor.

That allows to investigate networks in a number of ways, such as in one of their studies on citations, coauthorship, and clustering of scholars in different areas.

4/N
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 15d
Papalexakis then uses the example of a graph in which multiple edges between the same vertices are possible, which leads to one adjacency matrix for each label and therefore to a “3-dimensional matrix”, or tensor, if we want all of those adjacency matrices as part of a same mathematical object.

3/N
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 15d
Preparing the ground for a future generalization from matrices to tensors, Papalexakis characterized rank-one matrices as those that can be obtained from the product of a column and a row. That entails a perspective of low-rank approximations as a sum of few such column-row products.

2/N
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 15d
Evangelos Papalexakis gave an interesting and accessible keynote about latent structure and tensor decomposition at #bracis2025

The identification of latent structure was motivated with the observation from Miller’s Law that we only keep “7 +/- 2” items in memory at any given time time.

1/N
1
thserra.bsky.social
Thiago Serra @thserra.bsky.social · 17d
Today I learned that there is a dataset about puns in Portuguese called Puntuguese 😂

… and also that much better results on this dataset will be in the proceedings of #bracis2025

The talk was presented by Jhúlia de Souza Leal’s master’s advisor on behalf of the authors