Wasserstein Motifs: Non-deterministic Alignment of Ecological Networks

Yifan Xu, Carlos A Taveras, Lydia Beaudrot, Cesar A Uribe

Confirmation: I have read and agree with the workshop's policy on behalf of myself and my co-authors.

Track: long paper (up to 10 pages)

Keywords: Network Alignment, Graph Matching, Ecological Network, Optimal Transport, AI for Ecology

TL;DR: We present Wasserstein Motifs, a flexible network alignment method that uses an optimal transport-inspired approach to identify shared functional roles across different ecological communities.

Abstract: We study the problem of ecological network (food web) alignment, where we seek to identify structural equivalences among species and uncover backbones of interactions that represent shared functional substructures. These fundamental properties reveal the functional relationships that sustain ecosystems, enabling more accurate predictions of biodiversity responses to environmental change. Existing methods are computationally expensive, not scalable, and hard to interpret ecologically. We provide a first rigorous formalization of food web alignment based on network motifs, and show existing methods popularized in the ecological community are equivalent to minimizing a Fused Gromov-Wasserstein-like cost functional, termed Wasserstein Motifs. Moreover, we propose an interpretable and provably correct algorithm that efficiently computes non-deterministic alignments between food webs by leveraging their representation as feature measure networks. As a byproduct, we introduce a novel approach for identifying non-deterministic backbones of interactions in food webs. Experiments on a continental-scale dataset of 129 Sub-Saharan African mammal food webs demonstrate significant gains in accuracy, scalability, and interpretability over state-of-the-art methods. Our results establish a principled bridge between ecological network science and optimal transport, opening avenues for the analysis of complex ecological structured data.

Anonymization: This submission has been anonymized for double-blind review via the removal of identifying information such as names, affiliations, and identifying URLs.

Presenter: ~Yifan_Xu24

Format: Maybe: the presenting author will attend in person, contingent on other factors that still need to be determined (e.g., visa, funding).

Funding: Yes, the presenting author of this submission falls under ICLR’s funding aims, and funding would significantly impact their ability to attend the workshop in person.

Submission Number: 21