Improving historical spelling normalization with bi-directional LSTMs and multi-task learning | Read Paper on Bytez