Title:Controllable Length Control Neural Encoder-Decoder via Reinforcement Learning

Abstract:Controlling output length in neural language generation is valuable in many scenarios, especially for the tasks that have length constraints. A model with stronger length control capacity can produce sentences with more specific length, however, it usually sacrifices semantic accuracy of the generated sentences. Here, we denote a concept of Controllable Length Control (CLC) for the trade-off between length control capacity and semantic accuracy of the language generation model. More specifically, CLC is to alter length control capacity of the model so as to generate sentence with corresponding quality. This is meaningful in real applications when length control capacity and outputs quality are requested with different priorities, or to overcome unstability of length control during model training. In this paper, we propose two reinforcement learning (RL) methods to adjust the trade-off between length control capacity and semantic accuracy of length control models. Results show that our RL methods improve scores across a wide range of target lengths and achieve the goal of CLC. Additionally, two models LenMC and LenLInit modified on previous length-control models are proposed to obtain better performance in summarization task while still maintain the ability to control length.