MILE: Mixture of Incremental LoRA Experts for Continual Semantic Segmentation across Domains and Modalities
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915980800688128 |
|---|---|
| author | Muralidhara, Shishir Stricker, Didier Schuster, René |
| author_facet | Muralidhara, Shishir Stricker, Didier Schuster, René |
| contents | Continual semantic segmentation requires models to adapt to new domains or modalities without sacrificing performance on previously learned tasks. Expert-based learning, in which task-specific modules specialize in different domains, has proven effective in mitigating forgetting. These methods include dynamic expansion, which suffers from scalability issues, or parameter isolation, which constrains the ability to learn new tasks. We introduce Mixture of Incremental LoRA Experts (MILE), a modular and parameter-efficient framework for continual segmentation across both domains and modalities. MILE leverages Low-Rank Adaptation (LoRA) to instantiate lightweight experts for each new task while keeping the pretrained base network frozen. Each expert is trained exclusively on its task data, thus avoids overwriting previously learned information. A prototype-guided gating mechanism dynamically selects the most appropriate expert at inference. MILE achieves the benefits of expert-based learning while overcoming its scalability limitations. It requires only a marginal parameter increase per task and tens of LoRA adapters are needed before matching the size of a single full model, making it highly efficient in both training and storage. Across domain- and modality-incremental benchmarks, MILE achieves strong performance while ensuring better stability, plasticity, and scalability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_03555 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MILE: Mixture of Incremental LoRA Experts for Continual Semantic Segmentation across Domains and Modalities Muralidhara, Shishir Stricker, Didier Schuster, René Computer Vision and Pattern Recognition Continual semantic segmentation requires models to adapt to new domains or modalities without sacrificing performance on previously learned tasks. Expert-based learning, in which task-specific modules specialize in different domains, has proven effective in mitigating forgetting. These methods include dynamic expansion, which suffers from scalability issues, or parameter isolation, which constrains the ability to learn new tasks. We introduce Mixture of Incremental LoRA Experts (MILE), a modular and parameter-efficient framework for continual segmentation across both domains and modalities. MILE leverages Low-Rank Adaptation (LoRA) to instantiate lightweight experts for each new task while keeping the pretrained base network frozen. Each expert is trained exclusively on its task data, thus avoids overwriting previously learned information. A prototype-guided gating mechanism dynamically selects the most appropriate expert at inference. MILE achieves the benefits of expert-based learning while overcoming its scalability limitations. It requires only a marginal parameter increase per task and tens of LoRA adapters are needed before matching the size of a single full model, making it highly efficient in both training and storage. Across domain- and modality-incremental benchmarks, MILE achieves strong performance while ensuring better stability, plasticity, and scalability. |
| title | MILE: Mixture of Incremental LoRA Experts for Continual Semantic Segmentation across Domains and Modalities |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.03555 |