A dynamic multimodal network whose per-sample depth and parameters are set by prototype-based quality estimates, reporting large gains over prior trustworthy classifiers.
Predictive dynamic fusion,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Multi-QuAD: Multi-Level Quality-Adaptive Dynamic Network for Reliable Multimodal Classification
A dynamic multimodal network whose per-sample depth and parameters are set by prototype-based quality estimates, reporting large gains over prior trustworthy classifiers.