Scholar
Artur Back de Luca
Google Scholar ID: tL9d0UoAAAAJ
David R. Cheriton School of Computer Science, University of Waterloo
Citations & Impact
Citations
74
H-index
3
i10-index
2
Publications
6
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- 1. Learning to Add, Multiply, and Execute Algorithmic Instructions Exactly with Neural Networks (NeurIPS, 2025)
- 2. Exact Learning of Permutations for Nonzero Binary Inputs with Logarithmic Training Size and Quadratic Ensemble Complexity (High-dimensional Learning Dynamics Workshop @ ICML, 2025)
- 3. Positional Attention: Expressivity and Learnability of Algorithmic Computation (ICML, 2025)
- 4. Simulation of Graph Algorithms with Looped Transformers (ICML, 2024)
- 5. Local Graph Clustering with Noisy Labels (ICLR, 2024)
- 6. Mitigating Data Heterogeneity in Federated Learning with Data Augmentation (Preprint, 2022)
Research Experience
- Summer 2025: Applied Scientist Intern at Amazon New York, SCOT team, working with Ruijun Ma and Youxin Zhang on inbound event forecasting. 2022: Intern at Huawei’s Noah’s Ark Lab, focusing on Federated Learning and Domain Generalization with Guojun Zhang, and exploring invariant graph representations with Yingxue Zhang.
Education
- Ph.D. in Computer Science, University of Waterloo, Advisor: Kimon Fountoulakis; M.Sc. in AI and Robotics, Sapienza University of Rome; B.Sc. in Mechanical Engineering, Federal University of Santa Catarina.
Background
- Research Interests: Reasoning in neural networks, particularly their ability to learn algorithmic tasks; theoretical machine learning and graphs.
Co-authors