## An Introduction to Synchronous Grammars (2006)

Citations: | 12 - 0 self |

### BibTeX

@MISC{Chiang06anintroduction,

author = {David Chiang},

title = {An Introduction to Synchronous Grammars},

year = {2006}

}

### OpenURL

### Abstract

Synchronous context-free grammars are a generalization of context-free grammars (CFGs) that generate

### Citations

651 | Synchronous Tree Adjoining Grammars
- Sheiber, Schabes
- 1990
(Show Context)
Citation Context ... more powerful formalism called synchronous tree-adjoining grammar (TAG), first defined by Shieber and Schabes [17], but note that the 1994 formulation [16] supersedes it. In a tree-adjoining grammar =-=[7, 8]-=-, the elementary trees come in two flavors, initial trees, which are just as in TSG, and auxiliary trees, which have a leaf node called the foot node marked with an asterisk (∗). Here 9sis an example ... |

457 | Stochastic inversion transduction grammars and bilingual parsing of parallel corpora
- Wu
- 1997
(Show Context)
Citation Context ...tionship between two languages. Originally, they were developed in the late 1960s for programming-language compilation [1]. In natural language processing, they have been used for machine translation =-=[19, 20, 3]-=- and (less commonly, perhaps) semantic interpretation. As a preview, consider the following English sentence and its (admittedly somewhat unnatural) equivalent in Japanese (with English glosses): (1) ... |

381 | A hierarchical phrase-based model for statistical machine translation
- Chiang
- 2005
(Show Context)
Citation Context ...tionship between two languages. Originally, they were developed in the late 1960s for programming-language compilation [1]. In natural language processing, they have been used for machine translation =-=[19, 20, 3]-=- and (less commonly, perhaps) semantic interpretation. As a preview, consider the following English sentence and its (admittedly somewhat unnatural) equivalent in Japanese (with English glosses): (1) ... |

322 |
Tree adjunct grammars
- Joshi, Levy, et al.
- 1975
(Show Context)
Citation Context ... more powerful formalism called synchronous tree-adjoining grammar (TAG), first defined by Shieber and Schabes [17], but note that the 1994 formulation [16] supersedes it. In a tree-adjoining grammar =-=[7, 8]-=-, the elementary trees come in two flavors, initial trees, which are just as in TSG, and auxiliary trees, which have a leaf node called the foot node marked with an asterisk (∗). Here 9sis an example ... |

266 | A syntax-based statistical translation model
- Yamada, Knight
- 2001
(Show Context)
Citation Context ...tionship between two languages. Originally, they were developed in the late 1960s for programming-language compilation [1]. In natural language processing, they have been used for machine translation =-=[19, 20, 3]-=- and (less commonly, perhaps) semantic interpretation. As a preview, consider the following English sentence and its (admittedly somewhat unnatural) equivalent in Japanese (with English glosses): (1) ... |

234 | What’s in a translation rule
- Galley, Hopkins, et al.
- 2004
(Show Context)
Citation Context ...cussed by Rambow and Satta [13]. Linear nondeleting tree transducers Two syntax-based machine translation systems at ISI, that of Yamada and Knight [20] and another based on the work of Galley et al. =-=[5]-=-, are based on tree-to-string transducers [9]. The transducers they use are both linear and nondeleting, making them effectively equivalent to synchronous grammars. The Galley et al. formalism is equi... |

169 | Principles and implementation of deductive parsing
- Shieber, Schabes, et al.
- 1995
(Show Context)
Citation Context ...a pair of sentences and asked to find parse trees for both. Surprisingly, perhaps, the latter problem, where we are given more information, is in general much more difficult. Following Shieber et al. =-=[18]-=-, we present each parsing algorithm as a deductive system, which consists of a set of axioms that are taken as the items that are trivially provable, and a set of inference rules that say what items a... |

159 |
Mathematical and Computational Aspects of Lexicalized Grammars
- Schabes
- 1990
(Show Context)
Citation Context ...n both the source and target sides. The items should be proved in order of increasing j − i + j ′ − i ′ . 5 Extensions 5.1 Synchronous tree-substitution grammar Synchronous tree-substitution grammars =-=[15, 4]-=- are able to perform subject-object swapping as described above without flattening trees. In a (non-synchronous) tree-substitution grammar (TSG), the productions are elementary trees, tree fragments w... |

73 |
Syntax-directed transduction
- Lewis, Stearns
- 1968
(Show Context)
Citation Context ...form the kind of reordering seen here. But a synchronous CFG can do this. The term synchronous CFG is recent and far from universal. They were originally known as syntaxdirected transduction grammars =-=[10]-=- or syntax-directed translation schemata [1], the latter still probably being the most common name. In the NLP community they are also known as 2-multitext grammars [11], and inversion transduction gr... |

72 |
Syntax directed translations and the pushdown assembler
- Aho, Ullman
- 1969
(Show Context)
Citation Context ...ey are useful in many situations where one might want to specify a recursive relationship between two languages. Originally, they were developed in the late 1960s for programming-language compilation =-=[1]-=-. In natural language processing, they have been used for machine translation [19, 20, 3] and (less commonly, perhaps) semantic interpretation. As a preview, consider the following English sentence an... |

65 |
An overview of probabilistic tree transducers for natural language processing
- Knight, Graehl
- 2005
(Show Context)
Citation Context ...leting tree transducers Two syntax-based machine translation systems at ISI, that of Yamada and Knight [20] and another based on the work of Galley et al. [5], are based on tree-to-string transducers =-=[9]-=-. The transducers they use are both linear and nondeleting, making them effectively equivalent to synchronous grammars. The Galley et al. formalism is equivalent to a synchronous grammar that has a CF... |

50 | Multitext grammars and synchronous parsers
- Melamed
- 2003
(Show Context)
Citation Context ...directed transduction grammars [10] or syntax-directed translation schemata [1], the latter still probably being the most common name. In the NLP community they are also known as 2-multitext grammars =-=[11]-=-, and inversion transduction grammars [19] are a special case of synchronous CFGs. 2 Definition We give only an informal definition here. A synchronous CFG is like a CFG, but its productions have two ... |

40 |
Independent parallelism in finite copying parallel rewriting systems
- Rambow, Satta
- 1999
(Show Context)
Citation Context ...ynchronous TAGs. They are equivalent to simple range concatenation transducers [2] and would also be equivalent to the synchronous versions of a wide range of formalisms discussed by Rambow and Satta =-=[13]-=-. Linear nondeleting tree transducers Two syntax-based machine translation systems at ISI, that of Yamada and Knight [20] and another based on the work of Galley et al. [5], are based on tree-to-strin... |

34 |
Learning non-isomorphic tree mappings for machine translation
- Eisner
- 2003
(Show Context)
Citation Context ...n both the source and target sides. The items should be proved in order of increasing j − i + j ′ − i ′ . 5 Extensions 5.1 Synchronous tree-substitution grammar Synchronous tree-substitution grammars =-=[15, 4]-=- are able to perform subject-object swapping as described above without flattening trees. In a (non-synchronous) tree-substitution grammar (TSG), the productions are elementary trees, tree fragments w... |

30 | Generalized multitext grammars
- Melamed, Satta, et al.
- 2004
(Show Context)
Citation Context ...e of this notation is that you can compactly draw a synchronous derivation using a single tree. MTGs can also be extended to allow constituents to have an arbitrary number of discontiguous components =-=[12]-=-. The notation used in this extension, called generalized MTG, is much more closely related to the notation we have used here for synchronous CFG. Generalized MTGs also can simulate synchronous TAGs. ... |

23 | Some computational complexity results for synchronous context-free grammars
- Satta, Peserico
- 2005
(Show Context)
Citation Context ...f a rank-two synchronous CFG (or equivalent), bitext parsing takes O(n 6 ) time [19]. For a fixed synchronous CFG, bitext parsing takes O(n k ) time where k depends on the rank of the synchronous CFG =-=[14]-=-, but there is no k such that we can do bitext-parsing on an arbitrary synchronous CFG in O(n k ) time. 1 We present here Wu’s bitext-parsing algorithm for the rank two case. Assume that each right-ha... |

16 |
Restricting the weak generative capacity of synchronous tree adjoining grammar
- Shieber
- 1985
(Show Context)
Citation Context ... the power of synchronous CFG and TSG. We need a more powerful formalism called synchronous tree-adjoining grammar (TAG), first defined by Shieber and Schabes [17], but note that the 1994 formulation =-=[16]-=- supersedes it. In a tree-adjoining grammar [7, 8], the elementary trees come in two flavors, initial trees, which are just as in TSG, and auxiliary trees, which have a leaf node called the foot node ... |

9 | Factoring synchronous grammars by sorting
- Gildea, Satta, et al.
- 2006
(Show Context)
Citation Context ...G = 3-SCFG � 4-SCFG � . . . despite the fact that non-synchronous CFGs of rank 2 and higher are all weakly equivalent [1]. There is an efficient algorithm for minimizing the rank of a synchronous CFG =-=[6]-=-. No raising or lowering. The properties above considered synchronous CFGs as defining relations on strings. If you think of a synchronous CFG as defining a relation on trees, then all synchronous CFG... |

4 |
On the complexity of some extensions of RCG parsing
- Bertsch, Nederhof
(Show Context)
Citation Context ...an n-ary relation on strings. The two-dimensional case of multitext grammars is equivalent to synchronous CFGs. The following 2-multitext grammar rule is equivalent to our production (21): (49) S →⊲⊳ =-=[1, 2, 3]-=- [4, 2, 3, 1] � NP misses ɛ NP NP manque à NP Note that the symbols inside the round parentheses don’t appear in the order they will end up in; it is the numbers in square brackets, which correspond t... |