On the Correspondence between Compositional Matrix-Space Models of Language and Weighted Automata

From International Center for Computational Logic

Toggle side column

On the Correspondence between Compositional Matrix-Space Models of Language and Weighted Automata

Shima AsaadiShima Asaadi,  Sebastian RudolphSebastian Rudolph
Shima Asaadi, Sebastian Rudolph
On the Correspondence between Compositional Matrix-Space Models of Language and Weighted Automata
Proceedings of the ACL Workshop on Statistical Natural Language Processing and Weighted Automata (StatFSM 2016), August 2016
  • KurzfassungAbstract
    Compositional matrix-space models of language were recently proposed for the task of meaning representation of complex text structures in natural language processing. These models have been shown to be a theoretically elegant way to model compositionality in natural language. However, in practical cases, appropriate methods are required to learn such models by automatically acquiring the necessary token-to-matrix assignments. In this paper, we introduce graded matrix grammars of natural language, a variant of the matrix grammars proposed by Rudolph and Giesbrecht (2010), and show a close correspondence between this matrix-space model and weighted finite automata. We conclude that the problem of learning compositional matrix-space models can be mapped to the problem of learning weighted finite automata over the real numbers.
  • Projekt:Project: QuantLA
  • Forschungsgruppe:Research Group: Computational LogicComputational Logic
@inproceedings{AR2016,
  author    = {Shima Asaadi and Sebastian Rudolph },
  title     = {On the Correspondence between Compositional Matrix-Space Models of Language andWeighted Automata},
  booktitle = {Proceedings of the ACL Workshop on Statistical Natural Language Processing and Weighted Automata (StatFSM 2016)},
  year      = {2016},
  month     = {August}
}