Latent World Recovery for Multimodal Learning with Missing Modalities
English summary
The paper proposes Latent World Recovery (LWR), a framework for multimodal learning under missing modalities. LWR aligns modality-specific embeddings in a shared latent space and fuses only the modalities available at inference time to construct a unified representation, avoiding explicit imputation of missing data. Each modality is treated as a partial observation of an underlying latent state, and availability-aware representation learning is performed directly from observed modalities. Evaluated on real-world incomplete multi-omics benchmarks, LWR shows effective performance on cancer phenotype classification and survival prediction, addressing bioscience scenarios where not all modalities are always present.
Chinese summary
本文提出了潜在世界恢复（LWR）框架，用于缺失模态下的多模态学习。LWR在共享潜在空间中对齐各模态特定嵌入，并仅融合推理时实际可用的模态来构建统一表示，避免了显式插补缺失数据。该框架将每个模态视为潜在状态的部分感知，并直接从观察到的模态进行可用性感知的表示学习。在真实的不完整多组学基准上评估，LWR在癌症表型分类和生存预测任务中表现有效，适用于并非所有模态都始终可用的生物科学场景。
Key points
Proposes Latent World Recovery (LWR), a framework for multimodal learning with missing modalities that avoids imputation.
提出潜在世界恢复（LWR）框架，无需插补即可处理缺失模态的多模态学习。
Aligns modality-specific embeddings in a shared latent space and fuses only available modalities at inference time.
在共享潜在空间中对齐模态嵌入，并仅在推理时融合可用模态。
Treats each modality as a partial observation of a latent state, performing availability-aware representation learning.
将每个模态视为潜在状态的部分观测，进行可用性感知的表示学习。
Evaluated on real-world incomplete multi-omics data for cancer phenotype classification and survival prediction.
在真实不完整多组学数据上评估癌症表型分类和生存预测性能。