Adding a vaguely specified Bayesian component to a BiGRU-Transformer raises reported fake news test accuracy from 99.67% to 99.73% on one Kaggle dataset, with no code, data, or error bars.
Fake news classification based on content level features,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
A Hybrid Transformer Model for Fake News Detection: Leveraging Bayesian Optimization and Bidirectional Recurrent Unit
Adding a vaguely specified Bayesian component to a BiGRU-Transformer raises reported fake news test accuracy from 99.67% to 99.73% on one Kaggle dataset, with no code, data, or error bars.