Multimodal Generative Engine Optimization: Rank Manipulation for Vision-Language Model Rankers
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908774101417984 |
|---|---|
| author | Du, Yixuan Yu, Chenxiao Xu, Haoyan Wang, Ziyi Zhao, Yue Hu, Xiyang |
| author_facet | Du, Yixuan Yu, Chenxiao Xu, Haoyan Wang, Ziyi Zhao, Yue Hu, Xiyang |
| contents | Vision-Language Models (VLMs) are rapidly replacing unimodal encoders in modern retrieval and recommendation systems. While their capabilities are well-documented, their robustness against adversarial manipulation in competitive ranking scenarios remains largely unexplored. In this paper, we uncover a critical vulnerability in VLM-based product search: multimodal ranking attacks. We present Multimodal Generative Engine Optimization (MGEO), a novel adversarial framework that enables a malicious actor to unfairly promote a target product by jointly optimizing imperceptible image perturbations and fluent textual suffixes. Unlike existing attacks that treat modalities in isolation, MGEO employs an alternating gradient-based optimization strategy to exploit the deep cross-modal coupling within the VLM. Extensive experiments on real-world datasets using state-of-the-art models demonstrate that our coordinated attack significantly outperforms text-only and image-only baselines. These findings reveal that multimodal synergy, typically a strength of VLMs, can be weaponized to compromise the integrity of search rankings without triggering conventional content filters. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_12263 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Multimodal Generative Engine Optimization: Rank Manipulation for Vision-Language Model Rankers Du, Yixuan Yu, Chenxiao Xu, Haoyan Wang, Ziyi Zhao, Yue Hu, Xiyang Computation and Language Artificial Intelligence Machine Learning Vision-Language Models (VLMs) are rapidly replacing unimodal encoders in modern retrieval and recommendation systems. While their capabilities are well-documented, their robustness against adversarial manipulation in competitive ranking scenarios remains largely unexplored. In this paper, we uncover a critical vulnerability in VLM-based product search: multimodal ranking attacks. We present Multimodal Generative Engine Optimization (MGEO), a novel adversarial framework that enables a malicious actor to unfairly promote a target product by jointly optimizing imperceptible image perturbations and fluent textual suffixes. Unlike existing attacks that treat modalities in isolation, MGEO employs an alternating gradient-based optimization strategy to exploit the deep cross-modal coupling within the VLM. Extensive experiments on real-world datasets using state-of-the-art models demonstrate that our coordinated attack significantly outperforms text-only and image-only baselines. These findings reveal that multimodal synergy, typically a strength of VLMs, can be weaponized to compromise the integrity of search rankings without triggering conventional content filters. |
| title | Multimodal Generative Engine Optimization: Rank Manipulation for Vision-Language Model Rankers |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2601.12263 |