A lightweight estimator quantifies sample-level multimodal interactions (redundancy, uniqueness, synergy) in continuous distributions and uses them for data partitioning, distillation, and ensembling.
Baseline We adopt three primary types of multimodal learning paradigms: Feature-level fusion: Integration of multiple modalities at the feature level
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Efficient Quantification of Multimodal Interaction at Sample Level
A lightweight estimator quantifies sample-level multimodal interactions (redundancy, uniqueness, synergy) in continuous distributions and uses them for data partitioning, distillation, and ensembling.