Simpler Context-Dependent Logical Forms via Model Projections

Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pages 1456–1465, Berlin, Germany, August 7-12, 2016. c⃝2016 Association for Computational Linguistics Simpler Context-Dependent Logical Forms via Model Projections Reginald Long Stanford University reggylong@cs.stanford.edu Panupong Pasupat Stanford University ppasupat@cs.stanford.edu Percy Liang Stanford University pliang@cs.stanford.edu Abstract We consider the task of learning a context- dependent mapping from utterances to de- notations. With only denotations at train- ing time, we must search over a combina- torially large space of logical forms, which is even larger with context-dependent ut- terances. To cope with this challenge, we perform successive projections of the full model onto simpler models that operate over equivalence classes of logical forms. Though less expressive, we find that these simpler models are much faster and can be surprisingly effective. Moreover, they can be used to bootstrap the full model. Finally, we collected three new context- dependent semantic parsing datasets, and develop a new left-to-right parser. 1 Introduction Suppose we are only told that a piece of text (a command) in some context (state of the world) has some denotation (the effect of the command)—see Figure 1 for an example. How can we build a sys- tem to learn from examples like these with no ini- tial knowledge about what any of the words mean? We start with the classic paradigm of training semantic parsers that map utterances to logical forms, which are executed to produce the deno- tation (Zelle and Mooney, 1996; Zettlemoyer and Collins, 2005; Wong and Mooney, 2007; Zettle- moyer and Collins, 2009; Kwiatkowski et al., 2010). More recent work learns directly from de- notations (Clarke et al., 2010; Liang, 2013; Be- rant et al., 2013; Artzi and Zettlemoyer, 2013), but in this setting, a constant struggle is to con- tain the exponential explosion of possible logical forms. With no initial lexicon and longer context- dependent texts, our situation is exacerbated. Context: Text: Pour the last green beaker into beaker 2. Then into the first beaker. Mix it. Denotation: Figure 1: Our task is to learn to map a piece of text in some context to a denotation. An exam- ple from the ALCHEMY dataset is shown. In this paper, we ask: what intermediate logical form is suitable for modeling this mapping? In this paper, we propose projecting a full se- mantic parsing model onto simpler models over equivalence classes of logical form derivations. As illustrated in Figure 2, we consider the following sequence of models: • Model A: our full model that derives logi- cal forms (e.g., in Figure 1, the last utter- ance maps to mix(args[1][1])) compo- sitionally from the text so that spans of the ut- terance (e.g., “it”) align to parts of the logical form (e.g., args[1][1], which retrieves an argument from a previous logical form). This is based on standard semantic parsing (e.g., Zettlemoyer and Collins (2005)). • Model B: collapse all derivations with the same logical form; we map utterances to full logical forms, but without an alignment be- tween the …

Content truncated to 500 words