{"ID":6023503,"CreatedAt":"2026-07-08T01:00:23.257252134Z","UpdatedAt":"2026-07-10T10:09:03.016489495Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.06109","arxiv_id":"2607.06109","title":"RoME: Robust Mixture of Low-Rank Experts against Multiple Adversarial Perturbations","abstract":"Multi-perturbation adversarial training (MAT) aims to achieve robustness against multiple $\\ell_p$ perturbations but suffers from robustness trade-offs between different threats. To address this, we employ a mixture of experts (MoE) to route different threats through distinct model pathways. However, naive application of MoE encounters two critical challenges: experts tend to overlook threat-specific features and redundantly capture features shared across threats, and gating networks suffer from threat-agnostic routing where they learn nearly identical routing patterns across threats, thus preventing the construction of threat-specific model pathways. To this end, we propose Robust Mixture of Low-Rank Experts (RoME), where each expert is a low-rank additive update to the shared backbone, allowing it to capture threat-common features while experts focus on threat-specific information. To address threat-agnostic routing, RoME introduces (i) dual-scale gating that exploits threat-discriminative signals from local and global level features, and (ii) threat-guided gating diversification that enforces diverse expert utilization across threats. Extensive experiments demonstrate that RoME outperforms existing state-of-the-art MAT in union robustness and natural accuracy and improves robustness against unseen threats. Codes are available at https://github.com/wkim97/RoME.","short_abstract":"Multi-perturbation adversarial training (MAT) aims to achieve robustness against multiple $\\ell_p$ perturbations but suffers from robustness trade-offs between different threats. To address this, we employ a mixture of experts (MoE) to route different threats through distinct model pathways. However, naive application...","url_abs":"https://arxiv.org/abs/2607.06109","url_pdf":"https://arxiv.org/pdf/2607.06109v1","authors":"[\"Woo Jae Kim\",\"Kyle Min\",\"Suhyeon Ha\",\"Joonsung Jeon\",\"Sung-eui Yoon\"]","published":"2026-07-07T10:20:51Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.AI\"]","methods":"[\"Mixture of Experts\"]","has_code":false,"code_links":[{"ID":614022,"CreatedAt":"2026-07-08T01:00:23.257252134Z","UpdatedAt":"2026-07-08T01:00:23.257252134Z","DeletedAt":null,"paper_id":6023503,"paper_url":"https://arxiv.org/abs/2607.06109","paper_title":"RoME: Robust Mixture of Low-Rank Experts against Multiple Adversarial Perturbations","repo_url":"https://github.com/wkim97/RoME","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}