By decoupling text and image processing, refining features with mixtures of experts and AdaIN, and combining them through a confidence-based veto vote, GAMED reports state-of-the-art accuracy on Fakeddit (93.93%) and Yang (98.46%).
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
GAMED: Knowledge Adaptive Multi-Experts Decoupling for Multimodal Fake News Detection
By decoupling text and image processing, refining features with mixtures of experts and AdaIN, and combining them through a confidence-based veto vote, GAMED reports state-of-the-art accuracy on Fakeddit (93.93%) and Yang (98.46%).