Abstract
Semantic parsing is one of the key components of natural language understanding systems. A successful parse transforms an input utterance to an action that is easily understood by the system. Many algorithms have been proposed to solve this problem, from conventional rulebased or statistical slot-filling systems to shiftreduce based neural parsers. For complex parsing tasks, the state-of-the-art method is based on autoregressive sequence to sequence models to generate the parse directly. This model is slow at inference time, generating parses in O(n) decoding steps (n is the length of the target sequence). In addition, we demonstrate that this method performs poorly in zero-shot cross-lingual transfer learning settings. In this paper, we propose a non-autoregressive parser which is based on the insertion transformer to overcome these two issues. Our approach 1) speeds up decoding by 3x while outperforming the autoregressive model and 2) significantly improves cross-lingual transfer in the low-resource setting by 37% compared to autoregressive baseline. We test our approach on three well-known monolingual datasets: ATIS, SNIPS and TOP. For cross lingual semantic parsing, we use the MultiATIS++ and the multilingual TOP datasets.
Highlights
Given a query, a semantic parsing module identifies the intent of the query and extracts necessary slots that further refines the action to perform
We introduce a copy encoder outputs mechanism and achieve a significant improvement compared to the autoregressive decoder and sequence labeling on the zero-shot and fewshot setting in cross lingual transfer semantic parsing
Main Result: Table 1 shows the results from monolingual experiments on three datasets: Task Oriented Parsing (TOP), Airline Travel Information System (ATIS) and SNIPS
Summary
A semantic parsing module identifies the intent (play music, book a flight) of the query and extracts necessary slots (entities) that further refines the action to perform (which song to play? Where or when to go?). Gupta et al (2018) and Einolghozati et al (2019) propose to use a Shift-Reduce parser based on Recurrent Neural Network for these complex queries. Rongali et al (2020) propose directly generating the parse as a formatted sequence and design a unified model based on sequence to sequence generation and pointer networks. Their approach formulates the tagging problem into a generation task in which the target is constructed by combining all the necessary intents and slots in a flat sequence with no restriction on the semantic parse schema. Ht is a subsequence of the target sequence y that preserves order.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
