---
title: "Three Minds, One Student: Online Multi-Teacher Knowledge Distillation for Multimodal Recommenders"
authors:
  - "Hangtong Xu"
  - "Yuanbo Xu"
  - "En Wang"
doi: "10.24963/ijcai.2026/348"
url: "https://ijcai.org/proceedings/2026/348"
pdf: "https://ijcai.org/proceedings/2026/0348.pdf"
venue: "Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence"
year: 2026
pages: "3133-3141"
track: "Main Track"
keywords:
  - "Data Mining: Information retrieval"
  - "Data Mining: Collaborative filtering"
  - "Data Mining: Recommender systems"
---

# Three Minds, One Student: Online Multi-Teacher Knowledge Distillation for Multimodal Recommenders

Hangtong Xu, Yuanbo Xu, En Wang

*Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence*, pages 3133-3141. DOI: [10.24963/ijcai.2026/348](https://doi.org/10.24963/ijcai.2026/348)

## Abstract

Existing multimodal recommendation models using complex fusion mechanisms (e.g., attention) or multi-stage processes (e.g., early or late fusion) integrate different modalities. However, attention-based adaptive fusion is prone to shortcut learning, where dominant collaborative signals (ID) can overshadow other modalities. This dominance affects the entire fusion process: early fusion often amplifies biases driven by identity, while late fusion struggles to extract preference-relevant signals from misaligned modalities, even with alignment regularization. To address these issues, we propose Multi-Teacher Single-Student Online Distillation for Multimodal Recommendation (MTS2-4MM), which reframes the multimodal recommendation task from direct fusion to controllable knowledge transfer. Specifically, we construct multiple teachers to specialize in complementary perspectives, and a unified student distills their guidance via objectives at both the ranking and representation levels. This design explicitly controls modality contributions, improves robustness to modality noise and misalignment. Furthermore, we design a Modality-specific Preference Extractor to explicitly extract user preferences across different modalities equally. Extensive experiments across five real-world datasets demonstrate that MTS2-4MM consistently outperforms state-of-the-art baselines, achieving improvements of up to 7.22%.