Window-based Tagging

this project will need to use a framework that supports auto-diff (computation graph, backprop) in order to compute gradients, and implement a sequence tagger with window-based features.
its related to experiment with word pre-trained
embeddings, and with different word representations, on two different tagging tasks.