Researchers report a multimodal framework combining Bengali OCR, hybrid visual encoders, and vision-language reasoning that achieved 97% accuracy on a Bengali fake news dataset while generating ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results