Article Structure
Abstract
A central problem in grounded language acquisition is learning the correspondences between a rich world state and a stream of text which references that world state.
Introduction
Recent work in learning semantics has focused on mapping sentences to meaning representations (e.g., some logical form) given aligned sen-tence/meaning pairs as training data (Ge and Mooney, 2005; Zettlemoyer and Collins, 2005; Zettlemoyer and Collins, 2007; Lu et al., 2008).
Domains and Datasets
Our goal is to learn the correspondence between a text w and the world state s it describes.
Generative Model
To learn the correspondence between a text w and a world state s, we propose a generative model p(w | s) with latent variables specifying this correspondence.
Learning and Inference
Our learning and inference methodology is a fairly conventional application of Expectation Maximization (EM) and dynamic programming.
Experiments
Two important aspects of our model are the segmentation of the text and the modeling of the co-
Conclusion
We have presented a generative model of correspondences between a world state and an unsegmented stream of text.
Topics
generative model
- To deal with the high degree of ambiguity present in this setting, we present a generative model that simultaneously segments the text into utterances and maps each utterance to a meaning representation grounded in the world state.Page 1, “Abstract”
- To cope with these challenges, we propose a probabilistic generative model that treats text segmentation, fact identification, and alignment in a single unified framework.Page 1, “Introduction”
- To learn the correspondence between a text w and a world state s, we propose a generative model p(w | s) with latent variables specifying this correspondence.Page 3, “Generative Model”
- We used a simple generic model of rendering string fields: Let U) be a word chosen uniformly from those in v.Page 5, “Generative Model”
- We have presented a generative model of correspondences between a world state and an unsegmented stream of text.Page 8, “Conclusion”
See all papers in Proc. ACL 2009 that mention generative model.
See all papers in Proc. ACL that mention generative model.
Back to top.
language acquisition
- A central problem in grounded language acquisition is learning the correspondences between a rich world state and a stream of text which references that world state.Page 1, “Abstract”
- However, this degree of supervision is unrealistic for modeling human language acquisition and can be costly to obtain for building large-scale, broad-coverage language understanding systems.Page 1, “Introduction”
- A more flexible direction is grounded language acquisition : learning the meaning of sentences in the context of an observed world state.Page 1, “Introduction”
- 6It is interesting to note that this type of staged training is evocative of language acquisition in children: lexical associations are formed (Model 1) before higher-level discourse structure is learned (Model 3).Page 7, “Experiments”
See all papers in Proc. ACL 2009 that mention language acquisition.
See all papers in Proc. ACL that mention language acquisition.
Back to top.
meaning representation
- To deal with the high degree of ambiguity present in this setting, we present a generative model that simultaneously segments the text into utterances and maps each utterance to a meaning representation grounded in the world state.Page 1, “Abstract”
- Recent work in learning semantics has focused on mapping sentences to meaning representations (e.g., some logical form) given aligned sen-tence/meaning pairs as training data (Ge and Mooney, 2005; Zettlemoyer and Collins, 2005; Zettlemoyer and Collins, 2007; Lu et al., 2008).Page 1, “Introduction”
- In this less restricted data setting, we must resolve multiple ambiguities: (l) the segmentation of the text into utterances; (2) the identification of relevant facts, i.e., the choice of records and aspects of those records; and (3) the alignment of utterances to facts (facts are the meaning representations of the utterances).Page 1, “Introduction”
- Think of the words spanned by a record as constituting an utterance with a meaning representation given by the record and subset of fields chosen.Page 4, “Generative Model”
See all papers in Proc. ACL 2009 that mention meaning representation.
See all papers in Proc. ACL that mention meaning representation.
Back to top.
discourse structure
- ,rlrl), where each record 7“,- E s. This model is intended to capture two types of regularities in the discourse structure of language.Page 4, “Generative Model”
- 6It is interesting to note that this type of staged training is evocative of language acquisition in children: lexical associations are formed (Model 1) before higher-level discourse structure is learned (Model 3).Page 7, “Experiments”
- We did not experiment with Model 3 since the discourse structure on records in this domain is not at all governed by a simple Markov model on record types—indeed, most regions do not refer to any records at all.Page 8, “Experiments”
See all papers in Proc. ACL 2009 that mention discourse structure.
See all papers in Proc. ACL that mention discourse structure.
Back to top.
word alignment
- The alignment aspect of our model is similar to the HMM model for word alignment (Ney and Vogel, 1996).Page 3, “Generative Model”
- (2008) perform joint segmentation and word alignment for machine translation, but the nature of that task is different from ours.Page 3, “Generative Model”
- Many of the remaining errors are due to the garbage collection phenomenon familiar from word alignment models (Moore, 2004; Liang et al., 2006).Page 7, “Experiments”
See all papers in Proc. ACL 2009 that mention word alignment.
See all papers in Proc. ACL that mention word alignment.
Back to top.
