Tracing and Coordinating Cross-Layer Influence for Multimodal Model Merging

• URL: http://arxiv.org/abs/2609.12897v1
• Date: Fri, 11 Sep 2026 14:24:07 GMT
• Title: Tracing and Coordinating Cross-Layer Influence for Multimodal Model Merging
• Abstract summary: Multimodal model merging aims to consolidate task experts into a single model that retains their complementary capabilities.<n>Most unimodal model merging methods combine expert updates within individual layers, and multimodal approaches largely follow this design.<n>We propose TAC-Merge for tracing and coordinating cross-layer influence in multimodal model merging.
• Score: 6.356544634679287
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multimodal model merging aims to consolidate task experts into a single model that retains their complementary capabilities. Most unimodal model merging methods combine expert updates within individual layers, and multimodal approaches largely follow this design. However, an expert update changes the representations passed to subsequent layers, allowing its influence to propagate across depth and affect how visual and textual information interact. When visual and language updates are combined, later updates act on inputs already modified by earlier ones, coupling their effects. This poses two challenges: (1) how to characterize the multimodal influence of individual expert updates across depth, and (2) how to jointly combine expert updates based on their multimodal influence. To address these challenges, we propose TAC-Merge for tracing and coordinating cross-layer influence in multimodal model merging. It contains two modules, i.e., multimodal influence mapping (MIM) and coupled merge control (CMC). MIM constructs graphs of update effects and uses Ricci curvature together with expert predictions to define a shared fusion objective. CMC models interactions among coefficient adjustments and jointly optimizes regional weights to synthesize one shared model. Experiments across diverse multimodal tasks demonstrate the effectiveness of TAC-Merge in consolidating complementary expert capabilities and supporting generalization to unseen tasks.

Related papers

• PivotMerge: Bridging Heterogeneous Multimodal Pre-training via Post-Alignment Model Merging [80.17966517536102]
  We argue that the core of Multimodal Large Language Models pre-training lies in establishing effective cross-modal alignment.<n>Motivated by this insight, we introduce the post-alignment merging task, which aims to integrate cross-modal alignment capabilities learned from multimodal pre-training.<n>We propose PivotMerge, a post-alignment merging framework for cross-modal projectors.
  arXiv  Detail & Related papers  (2026-04-18T09:38:03Z)
• Reconstructing Content via Collaborative Attention to Improve Multimodal Embedding Quality [59.651410243721045]
  CoCoA is a Content reconstruction pre-training paradigm based on Collaborative Attention for multimodal embedding optimization.<n>We introduce an EOS-based reconstruction task, encouraging the model to reconstruct input from the corresponding EOS> embeddings.<n>Experiments on MMEB-V1 demonstrate that CoCoA built upon Qwen2-VL and Qwen2.5-VL significantly improves embedding quality.
  arXiv  Detail & Related papers  (2026-03-02T05:34:45Z)
• Cross-Modal Alignment via Variational Copula Modelling [54.25504956780864]
  It is essential to develop multimodal learning methods to aggregate various information from multiple modalities.<n>Existing methods mainly rely on concatenation or the Kronecker product, oversimplifying the interaction structure between modalities.<n>We propose a novel copula-driven multimodal learning framework, which focuses on learning the joint distribution of various modalities.
  arXiv  Detail & Related papers  (2025-11-05T05:28:28Z)
• Graph4MM: Weaving Multimodal Learning with Structural Information [52.16646463590474]
  Graphs provide powerful structural information for modeling intra- and inter-modal relationships.<n>Previous works fail to distinguish multi-hop neighbors and treat the graph as a standalone modality.<n>We propose Graph4MM, a graph-based multimodal learning framework.
  arXiv  Detail & Related papers  (2025-10-19T20:13:03Z)
• MANGO: Multimodal Attention-based Normalizing Flow Approach to Fusion Learning [12.821814562210632]
  This paper introduces a novel Multimodal Attention-based Normalizing Flow (MANGO) approach.<n>We propose a new Invertible Cross-Attention layer to develop the Normalizing Flow-based Model for multimodal data.<n>We also introduce three new cross-attention mechanisms: Modality-to-Modality Cross-Attention (MMCA), Inter-Modality Cross-Attention (IMCA), and Learnable Inter-Modality Cross-Attention (LICA)
  arXiv  Detail & Related papers  (2025-08-13T18:56:57Z)
• MultiSHAP: A Shapley-Based Framework for Explaining Cross-Modal Interactions in Multimodal AI Models [5.011371514152517]
  Multimodal AI models have achieved impressive performance in tasks that require integrating information from multiple modalities, such as vision and language.<n>How to explain cross-modal interactions in multimodal AI models remains a major challenge.
  arXiv  Detail & Related papers  (2025-08-01T12:19:18Z)
• Co-AttenDWG: Co-Attentive Dimension-Wise Gating and Expert Fusion for Multi-Modal Offensive Content Detection [0.0]
  Multi-modal learning has emerged as a crucial research direction.<n>Existing approaches often suffer from insufficient cross-modal interactions and rigid fusion strategies.<n>We propose Co-AttenDWG, co-attention with dimension-wise gating, and expert fusion.<n>We show that Co-AttenDWG achieves state-of-the-art performance and superior cross-modal alignment.
  arXiv  Detail & Related papers  (2025-05-25T07:26:00Z)
• Uni-MoE: Scaling Unified Multimodal LLMs with Mixture of Experts [54.529880848937104]
  We develop a unified MLLM with the MoE architecture, named Uni-MoE, that can handle a wide array of modalities. Specifically, it features modality-specific encoders with connectors for a unified multimodal representation. We evaluate the instruction-tuned Uni-MoE on a comprehensive set of multimodal datasets.
  arXiv  Detail & Related papers  (2024-05-18T12:16:01Z)
• NativE: Multi-modal Knowledge Graph Completion in the Wild [51.80447197290866]
  We propose a comprehensive framework NativE to achieve MMKGC in the wild. NativE proposes a relation-guided dual adaptive fusion module that enables adaptive fusion for any modalities. We construct a new benchmark called WildKGC with five datasets to evaluate our method.
  arXiv  Detail & Related papers  (2024-03-28T03:04:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.