High-order expansion of Neural Ordinary Differential Equations flows

Dario Izzo, Sebastien Origer, Giacomo Acciarini, and Francesco Biscani · 2025

Read High-order expansion of Neural Ordinary Differential Equations flows by Dario Izzo, Sebastien Origer, Giacomo Acciarini, and Francesco Biscani (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-2504-08769.txt.

Description

Dario Izzo and three co-authors build a math framework that opens up the black box of neural differential equations. Their 2025 paper treats models that learn a flow, a rule for how a system changes over time, directly from data. Such models are deep and flexible, but their learned rules are hard to inspect, which blocks trust in high-stakes uses.

Their tool, Event Transition Tensors, uses higher-order derivatives to describe exactly how the learned flow behaves near key events. Three demos show the range: tracking uncertainty in a data-driven prey-predator control model, reading a neural feedback controller, and mapping landing paths in a three-body system with a learned Hamiltonian. In each case the tensors turn vague behavior into explicit structure. The work gives these models a firmer theory base.

Catalog record

CREATORDario Izzo, Sebastien Origer, Giacomo Acciarini, and Francesco Biscani
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-08769

Browse the full library · About the Tower of Babel