Value-Agnostic Conversational Semantic Parsing

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 3666–3681 August 1–6, 2021. ©2021 Association for Computational Linguistics 3666 Value-Agnostic Conversational Semantic Parsing Emmanouil Antonios Platanios, Adam Pauls, Subhro Roy, Yuchen Zhang, Alex Kyte, Alan Guo, Sam Thomson, Jayant Krishnamurthy, Jason Wolfe, Jacob Andreas, Dan Klein Microsoft Semantic Machines sminfo@microsoft.com Abstract Conversational semantic parsers map user ut- terances to executable programs given dia- logue histories composed of previous utter- ances, programs, and system responses. Ex- isting parsers typically condition on rich repre- sentations of history that include the complete set of values and computations previously dis- cussed. We propose a model that abstracts over values to focus prediction on type- and function-level context. This approach provides a compact encoding of dialogue histories and predicted programs, improving generalization and computational efficiency. Our model in- corporates several other components, includ- ing an atomic span copy operation and struc- tural enforcement of well-formedness con- straints on predicted programs, that are par- ticularly advantageous in the low-data regime. Trained on the SMCALFLOW and TREEDST datasets, our model outperforms prior work by 7.3% and 10.6% respectively in terms of absolute accuracy. Trained on only a thou- sand examples from each dataset, it outper- forms strong baselines by 12.4% and 6.4%. These results indicate that simple representa- tions are key to effective generalization in con- versational semantic parsing. 1 Introduction Conversational semantic parsers, which translate natural language utterances into executable pro- grams while incorporating conversational context, play an increasingly central role in systems for interactive data analysis (Yu et al., 2019), instruc- tion following (Guu et al., 2017), and task-oriented dialogue (Zettlemoyer and Collins, 2009). An ex- ample of this task is shown in Figure 1. Typical models are based on an autoregressive sequence prediction approach, in which a detailed represen- tation of the dialogue history is concatenated to the input sequence, and predictors condition on this se- quence and all previously generated components of the output (Suhr et al., 2018). While this approach can capture arbitrary dependencies between inputs and outputs, it comes at the cost of sample- and computational inefficiency. We propose a new “value-agnostic” approach to contextual semantic parsing driven by type-based representations of the dialogue history and function- based representations of the generated programs. Types and functions have long served as a founda- tion for formal reasoning about programs, but their use in neural semantic parsing has been limited, e.g., to constraining the hypothesis space (Krishna- murthy et al., 2017), guiding data augmentation (Jia and Liang, 2016), and coarsening in coarse-to-fine models (Dong and Lapata, 2018). We show that representing conversation histories and partial pro- grams via the types and functions they contain en- ables fast, accurate, and sample-efficient contextual semantic parsing. We propose a neural encoder– decoder contextual semantic parsing model which, in contrast to prior work: 1. uses a compact yet informative representation of discourse context in the encoder that con- siders only the types of …

Content truncated to 500 words