Week 01 — Motivations — Why Geometry in Deep Learning

AlphaFold, AlphaGo, equivariant molecular models — none works without geometry-aware architecture. The unifying thesis of geometric deep learning, due to Bronstein et al. 2021.

Geom DL  ·  schedule  ·  Week 01 of 12 ·  02 →

Week 01 — Motivations — Why Geometry in Deep Learning

AlphaFold, AlphaGo, equivariant molecular models — none works without geometry-aware architecture. The unifying thesis of geometric deep learning, due to Bronstein et al. 2021.

Lecture

Felix Klein’s 1872 Erlangen Program · the Bronstein-et-al. unifying thesis (2021) · the five geometric priors (translation, rotation, permutation, gauge, scale) · why a CNN works and why an MLP-on-flattened-pixels doesn’t.

Read before the lecture

Reference text for this week: chapter 01 of the bilingual notes — EN PDF · FR PDF.