Parsing entire discourses as very long strings: Capturing topic continuity in grounded language learning

Parsing entire discourses as very long strings: Capturing topic continuity in grounded language learning Minh-Thang Luong Department of Computer Science Stanford University Stanford, California lmthang@stanford.edu Michael C. Frank Department of Psychology Stanford University Stanford, California mcfrank@stanford.edu Mark Johnson Department of Computing Macquarie University Sydney, Australia Mark.Johnson@MQ.edu.au Abstract Grounded language learning, the task of map- ping from natural language to a representation of meaning, has attracted more and more in- terest in recent years. In most work on this topic, however, utterances in a conversation are treated independently and discourse struc- ture information is largely ignored. In the context of language acquisition, this indepen- dence assumption discards cues that are im- portant to the learner, e.g., the fact that con- secutive utterances are likely to share the same referent (Frank et al., 2013). The current pa- per describes an approach to the problem of simultaneously modeling grounded language at the sentence and discourse levels. We com- bine ideas from parsing and grammar induc- tion to produce a parser that can handle long input strings with thousands of tokens, creat- ing parse trees that represent full discourses. By casting grounded language learning as a grammatical inference task, we use our parser to extend the work of Johnson et al. (2012), investigating the importance of discourse con- tinuity in children’s language acquisition and its interaction with social cues. Our model boosts performance in a language acquisition task and yields good discourse segmentations compared with human annotators. 1 Introduction Learning mappings between natural language (NL) and meaning representations (MR) is an important goal for both computational linguistics and cognitive science. Accurately learning novel mappings is cru- cial in grounded language understanding tasks and such systems can suggest insights into the nature of children language learning. Two influential examples of grounded language learning tasks are the sportscasting task, RoboCup, where the NL is the set of running commentary and the MR is the set of logical forms representing ac- tions like kicking or passing (Chen and Mooney, 2008), and the cross-situational word-learning task, where the NL is the caregiver’s utterances and the MR is the set of objects present in the context (Siskind, 1996; Yu and Ballard, 2007). Work in these domains suggests that, based on the co- occurrence between words and their referents in context, it is possible to learn mappings between NL and MR even under substantial ambiguity. Nevertheless, contexts like RoboCup—where ev- ery single utterance is grounded—are extremely rare. Much more common are cases where a sin- gle topic is introduced and then discussed at length throughout a discourse. In a television news show, for example, a topic might be introduced by present- ing a relevant picture or video clip. Once the topic is introduced, the anchors can discuss it by name or even using a pronoun without showing a picture. The discourse is grounded without having to ground every utterance. Moreover, although previous work has largely treated utterance order as independent, the order of utterances is critical in grounded discourse contexts: …

Content truncated to 500 words