Predicting Stance to Detect Misinformation in Few-shot Learning