Delexicalized Paraphrase Generation

Abstract

We present a neural model for paraphrasing and train it to generate delexicalized sentences. We achieve this by creating training data in which each input is paired with a number of reference paraphrases. These sets of reference paraphrases represent a weak type of semantic equivalence based on annotated slots and intents. To understand semantics from different types of slots, other than anonymizing slots, we apply convolutional neural networks (CNN) prior to pooling on slot values and use pointers to locate slots in the output. We show empirically that the generated paraphrases are of high quality, leading to an additional 1.29% exact match on live utterances. We also show that natural language understanding (NLU) tasks, such as intent classification and named entity recognition, can benefit from data augmentation using automatically generated paraphrases.

Addressing Slot-Value Changes in Task-oriented Dialogue Systems Through Dialogue Domain Adaptation

“…Depending on the complexity of the changes that occur, domain adaptation of data-driven systems can be approached in two directions: (i) improving the model robustness, and (ii) adapting the training dialogues to the new situation. While the first direction has been largely explored through several techniques, including transfer learning (Louvan and Magnini, 2019), and zero-shot learning though schema-guided models (Wu et al, 2019a;Kim et al, 2020;Zhang et al, 2019;Heck et al, 2020;, and delexicalization (Henderson et al, 2014a,b;Yu et al, 2020), in this paper we take the second, less investigated, perspective, focusing on the relation between training dialogues and domain knowledge.…”