# Beyond Modality Harmony: Orthogonal Purification and Topology-Guided MoE for Conflict-Aware Multimodal Recommendation

arXiv: https://arxiv.org/abs/2609.02152 · HTML: https://mlsanity.com/paper/2609.02152/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in MoE for recommender systems, one of 12 directions in Mixture-of-experts.
- Builds on 22 of its 39 references that are in the map.
- First posted on arXiv 2 Sep 2026.

## Placed in

- Mixture-of-experts / [MoE for recommender systems](https://mlsanity.com/moe/moe-for-recommender-systems/)

## Builds on (most cited first, 22 of 22)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer](https://mlsanity.com/paper/1701.06538/) (2017)
- [Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere](https://mlsanity.com/paper/2005.10242/) (2020)
- [LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation](https://mlsanity.com/paper/2002.02126/) (2020)
- [Image-based Recommendations on Styles and Substitutes](https://mlsanity.com/paper/1506.04757/) (2015)
- [Contrastive Learning with Hard Negative Samples](https://mlsanity.com/paper/2010.04592/) (2020)
- [Balanced Multimodal Learning via On-the-fly Gradient Modulation](https://mlsanity.com/paper/2203.15332/) (2022)
- [Are Graph Augmentations Necessary? Simple Graph Contrastive Learning for Recommendation](https://mlsanity.com/paper/2112.08679/) (2021)
- [VBPR: Visual Bayesian Personalized Ranking from Implicit Feedback](https://mlsanity.com/paper/1510.01784/) (2015)
- [Learning Disentangled Representations for Recommendation](https://mlsanity.com/paper/1910.14238/) (2019)
- [Mining Latent Structures for Multimedia Recommendation](https://mlsanity.com/paper/2104.09036/) (2021)
- [GRCN: Graph-Refined Convolutional Network for Multimedia Recommendation with Implicit Feedback](https://mlsanity.com/paper/2111.02036/) (2021)
- [Bootstrap Latent Representations for Multi-modal Recommendation](https://mlsanity.com/paper/2207.05969/) (2022)
- [A Tale of Two Graphs: Freezing and Denoising Graph Structures for Multimodal Recommendation](https://mlsanity.com/paper/2211.06924/) (2022)
- [Debiased Contrastive Learning for Sequential Recommendation](https://mlsanity.com/paper/2303.11780/) (2023)
- [LGMRec: Local and Global Graph Learning for Multimodal Recommendation](https://mlsanity.com/paper/2312.16400/) (2023)
- [DiffMM: Multi-Modal Diffusion Model for Recommendation](https://mlsanity.com/paper/2406.11781/) (2024)
- [MENTOR: Multi-level Self-supervised Learning for Multimodal Recommendation](https://mlsanity.com/paper/2402.19407/) (2024)
- [Improving Multi-modal Recommender Systems by Denoising and Aligning Multi-modal Content and User Feedback](https://mlsanity.com/paper/2406.12501/) (2024)
- [Cold-start Sequential Recommendation via Meta Learner](https://mlsanity.com/paper/2012.05462/) (2020)
- [Layer-refined Graph Convolutional Networks for Recommendation](https://mlsanity.com/paper/2207.11088/) (2022)

## Abstract (arXiv)

Multimodal Recommender Systems (MRSs) typically rely on a flawed "modality harmony" assumption, presuming that multimodal features are inherently beneficial and strictly aligned with users' collaborative interaction patterns. However, modality-topology conflicts are ubiquitous in real-world scenarios due to deceptive visual clickbaits and mismatched semantics. Blindly integrating these noisy modalities inevitably pollutes the pristine collaborative space, causing severe representation distortion. To address this, we propose Orthogonal purification and topology-guided MoE for conflict-aware multimodal Recommendation (OrthoRec). At its core, OrthoRec introduces Collaborative-Guided Orthogonal Purification (CGOP), which geometrically decouples multimodal features into directions parallel and orthogonal to a pure collaborative anchor. By adaptively truncating the orthogonal noise with an energy-preserving normalization, CGOP rectifies deceptive semantic directions while preserving the modality's intrinsic representation capacity. Furthermore, we design a Topology-Aware Routing Mixture-of-Experts (TAR-MoE). Guided by the collaborative topology, TAR-MoE employs decoupled sigmoid gating to break the zero-sum bottleneck of traditional softmax attention, autonomously determining the injection scale for each purified modality. Finally, a safe-SSL objective is introduced to dynamically penalize the forced contrastive alignment of contradictory pairs. Experiments on three real-world Amazon datasets show that OrthoRec consistently outperforms competitive recent baselines and exhibits improved robustness under modality noise and item sparsity.