Latent World Recovery for Multimodal Learning with Missing Modalities
2026-06-10 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors address the problem of making predictions when some types of data (modalities) are missing, which is common in bioscience. They propose a method called Latent World Recovery (LWR) that aligns different data types into a shared space and combines only the available data for each case. Instead of trying to guess the missing data, their approach learns from what's present to make reliable predictions. They tested LWR on real biological datasets and found it works well for tasks like classifying cancer types and predicting patient survival.
multimodal learningmissing modalitieslatent spaceembeddingmodalities fusionmulti-omicscancer phenotype classificationsurvival prediction
Authors
Hui Wang, Tianyu Ren, Joseph Butler, Christopher Baker, Karen Rafferty, Simon McDade
Abstract
We study multimodal learning under missing modalities, with particular motivation from bioscience applications in which heterogeneous modalities are often only partially available when decisions need to be made. We propose Latent World Recovery (LWR), a framework built on two key ideas: (i) modality-specific embeddings from different modalities are aligned in a shared latent space, and (ii) a unified representation is constructed by fusing only the embeddings of the modalities that are actually available at both training and inference time. Rather than imputing missing modalities or requiring a fixed modality set, LWR treats each modality as a partial perception of an underlying latent state and performs availability-aware representation learning directly from the observed modalities. This combination of neighbor-based latent alignment and availability-aware modality fusion enables robust multimodal prediction under partial observation, while avoiding error propagation from explicit reconstruction of missing modalities. We evaluate the proposed framework on real-world incomplete multi-omics benchmarks and demonstrate that it provides an effective approach to downstream tasks such as cancer phenotype classification and survival prediction.