Research Area:  Machine Learning

Abstract:

Multi-scenario text generation is an essential task in natural language generation because of the multi-scene interlaced property of real-world problems. Traditional methods typically train the multi-scenario text generation models based on maximum likelihood estimation, which may suffer from the problem of exposure bias. Reinforcement learning (RL) based text generation methods could mitigate the exposure bias problem to some extent. However, the RL-based text generation methods are limited to the single-scenario tasks, which cannot be straightforwardly generalized to new scenario tasks. To address this problem, in this paper, we propose a multi-scenario text generation method based on meta RL (MetaRL-TG), which implements the method of model-agnostic meta-learning (MAML) in the framework of RL-based text generation. The proposed MetaRL-TG method first learns the initial parameters from multiple training tasks, then fine-tunes them in the target task. Thus, the proposed method is expected to efficiently achieve high-quality generated text in the new scenario. Finally, the effectiveness and generalization capability of the proposed method are demonstrated for eight scenarios through English test datasets.

Keywords:  
text generation method
meta reinforcement learning

Author(s) Name:  Tingting Zhao,Guixi Li,Yajing Song,Yuan Wang,Yarui Chen,Jucheng Yang

Publisher name:  ACM Digital Library

DOI:  10.1016/j.patrec.2022.11.031

Volume Information:  Volume 165,Pages 47-54,

Paper Link:   https://dl.acm.org/doi/abs/10.1016/j.patrec.2022.11.031