Neural Query Auto Completion

AbstractCurrent neural query auto-completion (QAC) systems rely on character-level language models, but they slow down when queries are long. We present how to utilize subword language models for the fast …...