Modality-Aware Identity Construction and Counterfactual Structure Learning for ID-Free Multimodal Recommendation

2026-05-18

AI Summary

Existing ID-agnostic multimodal recommendation methods struggle to capture long-tail semantic relationships due to static identity representations, insufficient exploitation of multimodal semantics, and popularity bias contaminating graph learning. To address these limitations, this work proposes MAIL, a novel framework that dynamically integrates multimodal semantics with positional encodings through a modality-aware identity construction module, yielding content-aware identity representations. Furthermore, MAIL introduces a counterfactual reasoning–based graph structure learning mechanism that incorporates a popularity penalty to uncover semantically relevant yet low-exposure neighbors, thereby mitigating popularity bias. Extensive experiments on five Amazon datasets demonstrate that MAIL achieves significant improvements over state-of-the-art baselines, with average gains of 7.81% in Recall@10 and 12.81% in NDCG@10.

Abstract

Multimodal recommendation has attracted extensive attention by leveraging heterogeneous modality information to alleviate data sparsity and improve recommendation accuracy. Existing methods have attempted to replace ID embeddings with multimodal features and have achieved promising preliminary results. However, these methods still exhibit the following two limitations: (1) the reconstructed ID representations remain relatively static and fail to fully exploit multimodal semantics; and (2) the graph learning process is insufficient in mining latent long-tail semantic relations and is easily affected by popularity bias. To address these issues, we propose a novel method named Modality-Aware Identity Construction and Counterfactual Structure Learning for ID-free Multimodal Recommendation (MAIL). Specifically, we design a modality-aware identity construction module that dynamically modulates positional encodings with multimodal semantics to construct content-aware ID-free identity representations. Then, we propose a counterfactual structure learning paradigm that mines low-exposure semantic neighbors via popularity penalization and alleviates popularity bias. Extensive experiments are conducted on five public Amazon datasets. Experimental results show that MAIL achieves average improvements of 7.81% in Recall@10 and 12.81% in NDCG@10 compared with the baseline models. Our code is available at https://github.com/HubuKG/MAIL.

Problem

• multimodal recommendation
• ID-free representation
• popularity bias
• long-tail semantic relations
• modality semantics

Innovation

• modality-aware identity construction
• counterfactual structure learning
• ID-free recommendation
• popularity bias mitigation
• multimodal semantics