Look, Listen and Pay More Attention: Fusing Multi-Modal Information for Video Violence Detection.