Improving LSTM-based Video Description with Linguistic Knowledge Mined from Text | Read Paper on Bytez