Open-set Cross Modal Generalization via Multimodal Unified Representation
About
This paper extends Cross Modal Generalization (CMG) to open-set environments by proposing the more challenging Open-set Cross Modal Generalization (OSCMG) task. This task evaluates multimodal unified representations in open-set conditions, addressing the limitations of prior closed-set cross-modal evaluations. OSCMG requires not only cross-modal knowledge transfer but also robust generalization to unseen classes within new modalities, a scenario frequently encountered in real-world applications. Existing multimodal unified representation work lacks consideration for open-set environments. To tackle this, we propose MICU, comprising two key components: Fine-Coarse Masked multimodal InfoNCE (FCMI) and Cross modal Unified Jigsaw Puzzles (CUJP). FCMI enhances multimodal alignment by applying contrastive learning at both holistic semantic and temporal levels, incorporating masking to enhance generalization. CUJP enhances feature diversity and model uncertainty by integrating modality-agnostic feature selection with self-supervised learning, thereby strengthening the model's ability to handle unknown categories in open-set tasks. Extensive experiments on CMG and the newly proposed OSCMG validate the effectiveness of our approach. The code is available at https://github.com/haihuangcode/CMG.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Event Classification (A → V) | VGGSound-AVEL 90K | Precision67 | 15 | |
| Event Classification (V → A) | VGGSound-AVEL 40K | Precision75.3 | 15 | |
| Event Classification (A → V) | VGGSound-AVEL 40K | Precision57.1 | 11 | |
| Event Classification (V → A) | VGGSound-AVEL 90K | Precision43.1 | 11 | |
| Event Localization (A → V) | VGGSound-AVEL 40K | Segment-level Accuracy56.2 | 11 | |
| Event Localization (A → V) | VGGSound-AVEL 90K | Segment-level Accuracy47.1 | 11 | |
| Event Localization (V → A) | VGGSound-AVEL 40K | Segment-level Accuracy59.5 | 11 | |
| Event Localization (V → A) | VGGSound-AVEL 90K | Segment-level Accuracy45.6 | 11 | |
| Cross-modal retrieval | MSCOCO V↔T zero-shot | R@11.3 | 6 | |
| Cross-modal retrieval | Clotho A↔T zero-shot | Recall@13.35 | 6 |