Paper 2026/1142 Original: eprint.iacr.org/2026/1142 ↗

A computational framework for principally polarized abelian varieties and applications

Maria Corte-Real Santos
Etienne Piasecki
Benjamin Wesolowski
Abstract

We construct a new framework for cryptographers to work with principally polarized abelian varieties (PPAVs). This framework offers a computational approach to abelian varieties agnostic to the choice of a coordinate system, culminating in the definition of an efficient model for principally polarised abelian varieties. We exhibit an instantiation of our framework by means of the theta model, thereby streamlining the documented capacities of the model, and extending them with new fundamental algorithms, like the computation of automorphism groups. Our framework focuses on what can be done with these objects, computationally, while relegating low-level considerations to the background, like the specific choice of a coordinate system (and thus the necessity to rely on Mumford's theory of theta coordinates). We illustrate the utility of our framework by proving that we can interpolate polarised isogenies in any dimension, generalizing to higher dimensions the most disruptive algorithm for elliptic curves in recent years. We prove that this interpolation offers a universal, canonical, and compact way to represent isogenies.

Metadata
Available format(s)
PDF
Category
Public-key cryptography
Publication info
Preprint.
History
2026-06-02: last updated
2026-06-02: published
Short URL
https://ia.cr/2026/1142
License
https://creativecommons.org/licenses/by/4.0/

BibTeX

@misc{cryptoeprint:2026/1142,
      author = {Maria Corte-Real Santos and Etienne Piasecki and Benjamin Wesolowski},
      title = {A computational framework for principally polarized abelian varieties and applications},
      howpublished = {Cryptology {ePrint} Archive, Paper 2026/1142},
      year = {2026},
      url = {https://eprint.iacr.org/2026/1142}
}
Note: In order to protect the privacy of readers, this archive does not use cookies or embedded third party content.