Low-Resource Domain Adaptation for Compositional Task-Oriented Semantic Parsing

Abstract

Task-oriented semantic parsing is a critical component of virtual assistants, which is responsible for understanding the user's intents (set reminder, play music, etc.). Recent advances in deep learning have enabled several approaches to successfully parse more complex queries Rongali et al., 2020), but these models require a large amount of annotated training data to parse queries on new domains (e.g. reminder, music).In this paper, we focus on adapting taskoriented semantic parsers to low-resource domains, and propose a novel method that outperforms a supervised neural model at a 10-fold data reduction. In particular, we identify two fundamental factors for low-resource domain adaptation: better representation learning and better training techniques. Our representation learning uses BART (Lewis et al., 2020) to initialize our model which outperforms encoder-only pre-trained representations used in previous work. Furthermore, we train with optimization-based meta-learning (Finn et al., 2017) to improve generalization to lowresource domains. This approach significantly outperforms all baseline methods in the experiments on a newly collected multi-domain taskoriented semantic parsing dataset (TOPv2 1 ).

Span Pointer Networks for Non-Autoregressive Task-Oriented Semantic Parsing

“…We evaluate our approach on several task-oriented semantic parsing datasets. Notably, we bridge the quality gap between non-autogressive and autoregressive parsers, achieving 87 EM on TOPv2 (Chen et al, 2020). Furthermore, due to our more consistent gold frames, we show strong improvements in model generalization in both cross-domain and cross-lingual transfer in low-resource settings.…”

“…Task-oriented semantic parsing broadly consists of mapping textual utterances to structured frames Aghajanyan et al, 2020;Chen et al, 2020;Rongali et al, 2020;Li et al, 2021). Frames are structured semantic representations of utterances, and are comprised of both ontology tokens (e.g., intents and slots) and utterance tokens.…”

“…We experiment with the following task-oriented semantic parsing datasets: TOP (Task Oriented Pars-ing) , TOPv2 (Chen et al, 2020), and MTOP (Multilingual Task Oriented Parsing) (Li et al, 2021). TOP evaluates parsers' abilities to produce nested frames across the event and navigation domains.…”

“…We evaluate span pointer networks on three axes: quality, generalizability, and resources. First, we benchmark exact match (EM) on multiple task-oriented semantic parsing datasets, including TOPv2 (Chen et al, 2020) and TOP . Our non-autoregressive parser is competitive, achieving 87 EM on TOPv2, and matches or exceeds autoregressive parsers in many cases.…”

A Survey on Recent Approaches for Natural Language Processing in Low-Resource Scenarios

“…Given a set of auxiliary high-resource tasks and a low-resource target task, meta-learning trains a model to decide how to use the auxiliary tasks in the most beneficial way for the target task. For NLP, this approach has been evaluated on tasks such as sentiment analysis , user intent classification Chen et al, 2020b), natural language understanding (Dou et al, 2019), text classification (Bansal et al, 2020) and dialogue generation . Instead of having a set of tasks, Rahimi et al (2019) built an ensemble of language-specific NER models which are then weighted depending on the zero-or few-shot target language.…”