English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Multimodal learning significantly benefits cancer survival prediction,especially the integration of pathological images and genomic data. Despiteadvantages of multimodal learning for cancer survival prediction, massiveredundancy in multimodal data prevents it from extracting discriminative andcompact information: (1) An extensive amount of intra-modal task-unrelatedinformation blurs discriminability, especially for gigapixel whole slide images(WSIs) with many patches in pathology and thousands of pathways in genomicdata, leading to an ``intra-modal redundancy" issue. (2) Duplicated informationamong modalities dominates the representation of multimodal data, which makesmodality-specific information prone to being ignored, resulting in an``inter-modal redundancy" issue. To address these, we propose a new framework,Prototypical Information Bottlenecking and Disentangling (PIBD), consisting ofPrototypical Information Bottleneck (PIB) module for intra-modal redundancy andPrototypical Information Disentanglement (PID) module for inter-modalredundancy. Specifically, a variant of information bottleneck, PIB, is proposedto model prototypes approximating a bunch of instances for different risklevels, which can be used for selection of discriminative instances withinmodality. PID module decouples entangled multimodal data into compact distinctcomponents: modality-common and modality-specific knowledge, under the guidanceof the joint prototypical distribution. Extensive experiments on five cancerbenchmark datasets demonstrated our superiority over other methods.

Prototypical Information Bottlenecking and Disentangling for Multimodal Cancer Survival Prediction