Flopping for FLOPs: Leveraging Equivariance for Computational Efficiency

Georg Bökman, David Nordström, Fredrik Kahl · 2025

Read Flopping for FLOPs: Leveraging Equivariance for Computational Efficiency by Georg Bökman, David Nordström, Fredrik Kahl (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-05169.txt.

Description

Georg Bokman, David Nordstrom, and Fredrik Kahl wrote a machine learning paper in two thousand twenty-five on making symmetry-aware networks cheaper to run. Networks that respect mirror symmetry generalize better but usually cost more arithmetic per parameter. Their title plays on the fix: flopping, or mirroring, the image to save FLOPs, the unit of computing work.

The trick is to split features into mirror-symmetric and mirror-opposite parts. In that basis the network's linear layers break into two independent blocks, halving the arithmetic. The authors report cuts in both total operations and real running time, giving a practical route to scalable symmetry-aware designs.

Catalog record

CREATORGeorg Bökman, David Nordström, Fredrik Kahl
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-05169

Browse the full library · About the Tower of Babel