Controllable Meaning Representation to Text Generation: Linearization and Data Augmentation Strategies
Chris Kedzie, Kathleen R. McKeown · 2020
We study the degree to which neural sequenceto-sequence models exhibit fine-grained controllability when performing natural language generation from a meaning representation.Using two task-oriented dialogue generation benchmarks, we systematically compare the effect of four input linearization strategies on controllability and faithfulness.Additionally, we evaluate how a phrase-based data augmentation method can improve performance.We find that properly aligning input sequences during training leads to highly controllable generation, both when training from scratch or when fine-tuning a larger pre-trained model.Data augmentation further improves control on difficult, randomly generated utterance plans.