Tropical Backpropagation

E-print/Working paper (2023)

This work introduces tropicalization, a novel technique that delivers tropical neural networks as tropical limits of deep ReLU networks. Tropicalization transfers the initial weights from real numbers to ... [more ▼]

This work introduces tropicalization, a novel technique that delivers tropical neural networks as tropical limits of deep ReLU networks. Tropicalization transfers the initial weights from real numbers to those in the tropical semiring while maintain- ing the underlying graph of the network. After verifying that tropicalization will not affect the classification capacity of deep neural networks, this study introduces a tropical reformulation of backpropagation via tropical linear algebra. Tropical arithmetic replaces multiplication operations in the network with additions and addition operations with max, and therefore, theoretically, reduces the algorithmic complexity during the training and inference phase. We demonstrate the latter by simulating the tensor multiplication underlying the feed-forward process of state- of-the-art trained neural network architectures and compare the standard forward pass of the models with the tropical ones. Our benchmark results show that tropi- calization speeds up inference by 50 %. Hence, we conclude that tropicalization bears the potential to reduce the training times of large neural networks drastically. [less ▲]

Algebraic renormalization and Feynman integrals in configuration spaces

in Advances in Theoretical and Mathematical Physics (2014), 18(2), 469-511

Feynman integrals and motives of configuration spaces

in Communications in Mathematical Physics (2012), 313