Seminario: Optimización y Aprendizaje Automático (2026)

Viernes 13:00–14:30 | Salón 25

(FCEA en Gonzalo Ramirez)

Descripción

Este semestre el seminario tiene por objetivo discutir temas de investigación reciente en deep learning geométrico y teoría de aproximación.

Calendario del Seminario

Fecha Título Conferencista Links
20/3/2026 Reunión inicial — ver niveles de entrada? Organizadores Ver abajo
27/3/2026 Análisis de Fourier en grafos M. Fiori
3/4/2026 TURISMO
10/4/2026 Graph neural networks 1 P. Raigorodsky GNN1 (abajo)
17/4/2026 Graph Neural Networks 2 P. Raigorodsky GNN2 (abajo)
24/4/2026 El fenómeno del doble descenso. L. Frachelle DD1 (abajo)
1/5/2026 FESTIVO
8/5/2026 El fenómeno del doble descenso (parte 2) L. Frachelle DD2 (abajo)
15/5/2026 Optimal sampling for least‑squares approximation L. Bentancur CD Kernel (abajo)
22/5/2026 Machine learning y espacios de Barron M. Velasco Barron (abajo)
29/5/2026 TBA Harold
05/6/2026 TBA Harold
12/6/2026 TBA S. Lavayoll
19/6/2026 TBA B. Marenco

Referencias de interés

Título Autores Link Tema
Optimal sampling for least‑squares approximation Ben Adcock arXiv:2409.02342 CD kernel
CS4ML: A general framework for active learning with arbitrary data based on Christoffel functions Ben Adcock, Juan M. Cardenas, Nick Dexter arXiv:2306.00945 CD kernel
Geometry and Optimization of Shallow Polynomial Networks Y. Arjevani, J. Bruna, J. Kileel, E. Polak, M. Trager arXiv:2501.06074 DL / Optimization
Geometry of Polynomial Neural Networks Kaie Kubjas, Jiayi Li, Maximilian Wiesmann arXiv:2402.00949 DL / Algebraic Geometry
Convolutional Neural Networks on Manifolds: From Graphs and Back Zhiyang Wang, Luana Ruiz, Alejandro Ribeiro arXiv:2210.00376 Geometric DL / GNNs
The Barron Space and the Flow‑Induced Function Spaces for Neural Network Models Weinan E, Chao Ma, Lei Wu arXiv:1906.08039 Barron
Operator Learning of Lipschitz Operators: An Information-Theoretic Perspective Samuel Lanthaler arXiv:2406.18794 DL
Breaking the Curse of Dimensionality with Convex Neural Networks Francis Bach arXiv:1412.8690 Barron
Stability Properties of Graph Neural Networks Fernando Gama,Joan Bruna,Alejandro Ribeiro arXiv:1905.04497 GNN1
On Vanishing Gradients, Over-Smoothing, and Over-Squashing in GNNs: Bridging Recurrent and Graph Learning Álvaro Arroyo, Alessio Gravina, Benjamin Gutteridge, Federico Barbero, Claudio Gallicchio, Xiaowen Dong, Michael Bronstein, Pierre Vandergheynst arXiv:2502.10818 GNN2
Graph Attention Networks Petar Veličković, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Liò, Yoshua Bengio arXiv:1710.10903 GNN2
Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks Christopher Morris, Martin Ritzert, Matthias Fey, William L. Hamilton, Jan Eric Lenssen, Gaurav Rattan, Martin Grohe arXiv:1810.02244 GNN2
Reconciling Modern Machine Learning Practice and the Bias-Variance Trade-Off Mikhail Belkin, Daniel Hsu, Siyuan Ma, Soumik Mandal arXiv:1812.11118 DD1
Double Descent Demystified: Identifying, Interpreting & Ablating the Sources of a Deep Learning Puzzle Rylan Schaeffer, Mikail Khona, Zachary Robertson, Akhilan Boopathy, Kateryna Pistunova, Jason W. Rocks, Ila Fiete, Oluwasanmi Koyejo arXiv:2303.14151 DD1
Deep Double Descent: Where Bigger Models and More Data Hurt Preetum Nakkiran, Gal Kaplun, Dimitris Kalimeris, Tristan Yang, Benjamin L. Edelman, Fred Zhang, Boaz Barak arXiv:1912.02292 DD2
The Neural Tangent Kernel in High Dimensions: Triple Descent and a Multi-Scale Theory of Generalization Ben Adlam, Jeffrey Pennington PMLR v119 DD2

Organizadores