Methods, computer readable media, and devices for exceeding the limits of visual-linguistic
multi-task learning are disclosed. One method may include identifying a multi-modal multi …