TY  - RPRT
TI  - Deep Learning as a Convex Paradigm of Computation: Minimizing Circuit Size with ResNets
AU  - Arthur Jacot
PY  - 2026
UR  - https://arxiv.org/abs/2511.20888
ID  - 2511.20888
ER  -