Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387076306> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W4387076306 abstract "Predicting multiple real-world tasks in a single model often requires a particularly diverse feature space. Multimodal (MM) models aim to extract the synergistic predictive potential of multiple data types to create a shared feature space with aligned semantic meaning across inputs of drastically varying sizes (i.e. images, text, sound). Most current MM architectures fuse these representations in parallel, which not only limits their interpretability but also creates a dependency on modality availability. We present MultiModN, a multimodal, modular network that fuses latent representations in a sequence of any number, combination, or type of modality while providing granular real-time predictive feedback on any number or combination of predictive tasks. MultiModN's composable pipeline is interpretable-by-design, as well as innately multi-task and robust to the fundamental issue of biased missingness. We perform four experiments on several benchmark MM datasets across 10 real-world tasks (predicting medical diagnoses, academic performance, and weather), and show that MultiModN's sequential MM fusion does not compromise performance compared with a baseline of parallel fusion. By simulating the challenging bias of missing not-at-random (MNAR), this work shows that, contrary to MultiModN, parallel fusion baselines erroneously learn MNAR and suffer catastrophic failure when faced with different patterns of MNAR at inference. To the best of our knowledge, this is the first inherently MNAR-resistant approach to MM modeling. In conclusion, MultiModN provides granular insights, robustness, and flexibility without compromising performance." @default.
- W4387076306 created "2023-09-27" @default.
- W4387076306 creator A5005643082 @default.
- W4387076306 creator A5007940211 @default.
- W4387076306 creator A5018430480 @default.
- W4387076306 creator A5020200055 @default.
- W4387076306 creator A5035163294 @default.
- W4387076306 creator A5063796208 @default.
- W4387076306 creator A5073756389 @default.
- W4387076306 creator A5092950364 @default.
- W4387076306 date "2023-09-25" @default.
- W4387076306 modified "2023-09-28" @default.
- W4387076306 title "MultiModN- Multimodal, Multi-Task, Interpretable Modular Networks" @default.
- W4387076306 doi "https://doi.org/10.48550/arxiv.2309.14118" @default.
- W4387076306 hasPublicationYear "2023" @default.
- W4387076306 type Work @default.
- W4387076306 citedByCount "0" @default.
- W4387076306 crossrefType "posted-content" @default.
- W4387076306 hasAuthorship W4387076306A5005643082 @default.
- W4387076306 hasAuthorship W4387076306A5007940211 @default.
- W4387076306 hasAuthorship W4387076306A5018430480 @default.
- W4387076306 hasAuthorship W4387076306A5020200055 @default.
- W4387076306 hasAuthorship W4387076306A5035163294 @default.
- W4387076306 hasAuthorship W4387076306A5063796208 @default.
- W4387076306 hasAuthorship W4387076306A5073756389 @default.
- W4387076306 hasAuthorship W4387076306A5092950364 @default.
- W4387076306 hasBestOaLocation W43870763061 @default.
- W4387076306 hasConcept C101468663 @default.
- W4387076306 hasConcept C104317684 @default.
- W4387076306 hasConcept C111919701 @default.
- W4387076306 hasConcept C119857082 @default.
- W4387076306 hasConcept C138885662 @default.
- W4387076306 hasConcept C154945302 @default.
- W4387076306 hasConcept C162324750 @default.
- W4387076306 hasConcept C185592680 @default.
- W4387076306 hasConcept C187736073 @default.
- W4387076306 hasConcept C2776401178 @default.
- W4387076306 hasConcept C2780451532 @default.
- W4387076306 hasConcept C2781067378 @default.
- W4387076306 hasConcept C41008148 @default.
- W4387076306 hasConcept C41895202 @default.
- W4387076306 hasConcept C49937458 @default.
- W4387076306 hasConcept C55493867 @default.
- W4387076306 hasConcept C63479239 @default.
- W4387076306 hasConceptScore W4387076306C101468663 @default.
- W4387076306 hasConceptScore W4387076306C104317684 @default.
- W4387076306 hasConceptScore W4387076306C111919701 @default.
- W4387076306 hasConceptScore W4387076306C119857082 @default.
- W4387076306 hasConceptScore W4387076306C138885662 @default.
- W4387076306 hasConceptScore W4387076306C154945302 @default.
- W4387076306 hasConceptScore W4387076306C162324750 @default.
- W4387076306 hasConceptScore W4387076306C185592680 @default.
- W4387076306 hasConceptScore W4387076306C187736073 @default.
- W4387076306 hasConceptScore W4387076306C2776401178 @default.
- W4387076306 hasConceptScore W4387076306C2780451532 @default.
- W4387076306 hasConceptScore W4387076306C2781067378 @default.
- W4387076306 hasConceptScore W4387076306C41008148 @default.
- W4387076306 hasConceptScore W4387076306C41895202 @default.
- W4387076306 hasConceptScore W4387076306C49937458 @default.
- W4387076306 hasConceptScore W4387076306C55493867 @default.
- W4387076306 hasConceptScore W4387076306C63479239 @default.
- W4387076306 hasLocation W43870763061 @default.
- W4387076306 hasOpenAccess W4387076306 @default.
- W4387076306 hasPrimaryLocation W43870763061 @default.
- W4387076306 hasRelatedWork W3006943036 @default.
- W4387076306 hasRelatedWork W4200027074 @default.
- W4387076306 hasRelatedWork W4200511449 @default.
- W4387076306 hasRelatedWork W4206534706 @default.
- W4387076306 hasRelatedWork W4211177414 @default.
- W4387076306 hasRelatedWork W4229079080 @default.
- W4387076306 hasRelatedWork W4299487748 @default.
- W4387076306 hasRelatedWork W4385957992 @default.
- W4387076306 hasRelatedWork W4385965371 @default.
- W4387076306 hasRelatedWork W4386025632 @default.
- W4387076306 isParatext "false" @default.
- W4387076306 isRetracted "false" @default.
- W4387076306 workType "article" @default.