Simultaneous Translation with Flexible Policy via Restricted Imitation Learning

ACL 2019 Simultaneous translation is widely useful but remains one of the most difficult tasks in NLP. Previous work either uses fixed-latency policies, or train a complicated two-staged model using reinforcement learning. We propose a much simpler single model that adds a `delay' token to the target vocabulary, and design a restricted dynamic oracle to greatly simplify training. Experiments on ChineseEnglish simultaneous translation show that our work leads to flexible policies that achieve better BLEU scores and lower latencies compared to both fixed and RL-learned policies.