---
title: "LION: A Clifford Neural Paradigm for Multimodal-Attributed Graph Learning"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.24795v1"
permalink: "https://feed.astralor.com/items/itm_a069185bdadc6e8f3b7f2e05"
published: "2026-08-25T16:39:59.000Z"
score: 77
channel: "ai"
tags: ["arXiv","Clifford algebra","LION","Graph Neural Network","Multimodal-Attributed Graph"]
---

# LION: A Clifford Neural Paradigm for Multimodal-Attributed Graph Learning

## AI commentary

> The value of this paper lies in rethinking modality alignment and fusion from a geometric perspective rather than simply stacking network modules; it is highly inspiring for researchers working on graph foundation models and multimodal reasoning. If you are doing joint vision-language modeling, LION's propagation-then-aggregation design is especially worth studying.

## At a glance

Multimodal-attributed graph learning faces two major challenges: modality alignment that ignores graph context, and modality fusion that lacks adaptation. Existing methods mostly adopt topology-constrained or modality-specific operators as tokenizers, which hinder modality interaction and lead to suboptimal alignment. LION, based on Clifford algebra and a decoupled graph neural paradigm, first constructs a modality-aware geometric manifold to achieve modality alignment via high-order graph propagation. Then it proposes adaptive holographic aggregation, integrating component-wise energy and propagation-scale information with learnable parameters to improve modality fusion. Extensive experiments on 9 text-image multimodal-attributed graph datasets show that LION significantly outperforms state-of-the-art baselines across 3 graph and 3 modality downstream tasks.

## Key points

- Existing modal alignment methods neglect graph context; LION uses Clifford algebra to construct a geometric manifold for stronger modality interaction
- An adaptive holographic aggregation is proposed, integrating component-wise energy and propagation-scale information to improve modality fusion
- On 9 text-image multimodal-attributed graph datasets, LION surpasses SOTA baselines across 3 graph tasks and 3 modality tasks

## Full text

Recently, the rapid advancement of multimodal domains has driven a data-centric paradigm shift in graph ML, transitioning from text-attributed to multimodal-attributed graphs. This advancement significantly enhances data representation and expands the scope of graph downstream tasks, such as modality-oriented tasks, thereby improving the practical utility of graph ML. Despite its promise, limitations exist in the current neural paradigms:(1) Neglect Context in Modality Alignment: Most existing methods adopt topology-constrained or modality-specific operators as tokenizers.These aligners inevitably neglect graph context and inhibit modality interaction, resulting in suboptimal alignment.(2) Lack of Adaptation in Modality Fusion: Most existing methods are simple adaptations for 2-modality graphs and fail to adequately exploit aligned tokens equipped with topology priors during fusion, leading to poor generalizability and performance degradation.To address the above issues, we propose LION (c\underline{LI}ff\underline{O}rd \underline{N}eural paradigm) based on the Clifford algebra and decoupled graph neural paradigm (i.e., propagation-then-aggregation) to implement alignment-then-fusion in multimodal-attributed graphs. Specifically, we first construct a modality-aware geometric manifold grounded in Clifford algebra.This geometric-induced high-order graph propagation efficiently achieves modality interaction, facilitating modality alignment.Then, based on the topology-aware Clifford components of aligned tokens, we propose adaptive holographic aggregation. This module integrates component-wise energy and propagation-scale information with learnable parameters to improve modality fusion. Extensive experiments on 9 text-image MAG datasets demonstrate that LION significantly outperforms SOTA baselines across 3 graph and 3 modality downstream tasks.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_a069185bdadc6e8f3b7f2e05