Arie Soeteman
@ariesoeteman.bsky.social
PhD candidate in Machine Learning for Automated Reasoning @University of Amsterdam || Mathematical Logic || AI for ICUs in low-resource settings @GOAL3 || Yoga & Piano enthusiast
Today I will be presenting our presenting our poster at NeurIPS 🚀
Time: 4:30-7:30
Location: Exhibit Hall C,D,E
Drop by and say hello if you're interested in GNNs, modal logic, or isomorphism testing. See you this afternoon!
Time: 4:30-7:30
Location: Exhibit Hall C,D,E
Drop by and say hello if you're interested in GNNs, modal logic, or isomorphism testing. See you this afternoon!
December 3, 2025 at 3:28 PM
Today I will be presenting our presenting our poster at NeurIPS 🚀
Time: 4:30-7:30
Location: Exhibit Hall C,D,E
Drop by and say hello if you're interested in GNNs, modal logic, or isomorphism testing. See you this afternoon!
Time: 4:30-7:30
Location: Exhibit Hall C,D,E
Drop by and say hello if you're interested in GNNs, modal logic, or isomorphism testing. See you this afternoon!
How can we make graph neural networks more expressive?
We approach this via hybrid modal logics. As it turns out, the logical expressiveness of many GNNs out there reduces to that of the most common graph isomorphism solvers.
Interested? Read more on Arxiv: www.arxiv.org/abs/2506.13911
We approach this via hybrid modal logics. As it turns out, the logical expressiveness of many GNNs out there reduces to that of the most common graph isomorphism solvers.
Interested? Read more on Arxiv: www.arxiv.org/abs/2506.13911
Logical Expressiveness of Graph Neural Networks with Hierarchical Node Individualization
We propose and study Hierarchical Ego Graph Neural Networks (HEGNNs), an expressive extension of graph neural networks (GNNs) with hierarchical node individualization, inspired by the Individualizatio...
www.arxiv.org
June 20, 2025 at 9:44 AM
How can we make graph neural networks more expressive?
We approach this via hybrid modal logics. As it turns out, the logical expressiveness of many GNNs out there reduces to that of the most common graph isomorphism solvers.
Interested? Read more on Arxiv: www.arxiv.org/abs/2506.13911
We approach this via hybrid modal logics. As it turns out, the logical expressiveness of many GNNs out there reduces to that of the most common graph isomorphism solvers.
Interested? Read more on Arxiv: www.arxiv.org/abs/2506.13911
Theorem Proving and Machine Learning in the Age of LLMs —2025 was a remarkable event with stimulating people.
My main takeaway is that to solve frontier math problems arxiv.org/abs/2408.15332
we need ML for intermediate goal setting and abstraction, grounded in formal verification.
My main takeaway is that to solve frontier math problems arxiv.org/abs/2408.15332
we need ML for intermediate goal setting and abstraction, grounded in formal verification.
What makes math problems hard for reinforcement learning: a case study
Using a long-standing conjecture from combinatorial group theory, we explore, from multiple perspectives, the challenges of finding rare instances carrying disproportionately high rewards. Based on le...
arxiv.org
April 8, 2025 at 5:39 PM
Theorem Proving and Machine Learning in the Age of LLMs —2025 was a remarkable event with stimulating people.
My main takeaway is that to solve frontier math problems arxiv.org/abs/2408.15332
we need ML for intermediate goal setting and abstraction, grounded in formal verification.
My main takeaway is that to solve frontier math problems arxiv.org/abs/2408.15332
we need ML for intermediate goal setting and abstraction, grounded in formal verification.
Paper out; A step towards efficient tensor-network simulation of noisy quantum circuits.
Our method (1) allows for simulation under generic noise, (2) works in the low-noise regime, and (3) retains tensor network scaling, enabling simulations of large numbers of qubits.
arxiv.org/html/2501.13...
Our method (1) allows for simulation under generic noise, (2) works in the low-noise regime, and (3) retains tensor network scaling, enabling simulations of large numbers of qubits.
arxiv.org/html/2501.13...
Non-zero noise extrapolation: accurately simulating noisy quantum circuits with tensor networks
arxiv.org
January 25, 2025 at 12:07 PM
Paper out; A step towards efficient tensor-network simulation of noisy quantum circuits.
Our method (1) allows for simulation under generic noise, (2) works in the low-noise regime, and (3) retains tensor network scaling, enabling simulations of large numbers of qubits.
arxiv.org/html/2501.13...
Our method (1) allows for simulation under generic noise, (2) works in the low-noise regime, and (3) retains tensor network scaling, enabling simulations of large numbers of qubits.
arxiv.org/html/2501.13...