Low Resource Style Transfer via Domain Adaptive Meta Learning
This addresses the challenge of performance degradation and data requirements in text style transfer for low-resource domains, representing a novel method for a known bottleneck.
The paper tackles the problem of text style transfer in low-resource domains by proposing DAML-ATM, which combines domain-adaptive meta-learning and an adversarial transfer model to adapt to unseen domains with minimal data, achieving state-of-the-art results against ten baselines.
Text style transfer (TST) without parallel data has achieved some practical success. However, most of the existing unsupervised text style transfer methods suffer from (i) requiring massive amounts of non-parallel data to guide transferring different text styles. (ii) colossal performance degradation when fine-tuning the model in new domains. In this work, we propose DAML-ATM (Domain Adaptive Meta-Learning with Adversarial Transfer Model), which consists of two parts: DAML and ATM. DAML is a domain adaptive meta-learning approach to learn general knowledge in multiple heterogeneous source domains, capable of adapting to new unseen domains with a small amount of data. Moreover, we propose a new unsupervised TST approach Adversarial Transfer Model (ATM), composed of a sequence-to-sequence pre-trained language model and uses adversarial style training for better content preservation and style transfer. Results on multi-domain datasets demonstrate that our approach generalizes well on unseen low-resource domains, achieving state-of-the-art results against ten strong baselines.