An extension on the Hierarchical Recurrent Encoder-Decoder for Generative Context-Aware Query Suggestion, our implementation is in Tensorflow and uses an attention mechanism.

The original paper: A Hierarchical Recurrent Encoder-Decoder for Generative Context-Aware Query Suggestion

Our final report: HRED with Attention

The original Theano implementation: sordonia/hred-qs

• Data:
  • Raw AOL dataset
  • Chuncked pre-processed data
  • Data to feed into the model

Implementation by: Maartje ter Hoeve, Jörg Sander, Maurits Bleeker and Thijs Scheepers