Structured language modeling

By Ciprian Chelba, Frederick Jelinek. 2000

317 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

LSTMs Can Learn Syntax-Sensitive Dependencies Well, But Modeling Structure Makes Them Better (2018, neutral): “As RNNGs build phrase-structure trees through top-down operations, we propose extensions to the structure-building sequences and model architecture that enable left-corner (Henderson, 2003, 2004) and bottom-up (Chelba and Jelinek, 2000; Emami and Jelinek, 2005) generation orders (§4).”
TopicRNN: A Recurrent Neural Network with Long-Range Semantic Dependency (2016, background): “Previous work by Chelba and Jelinek (2000) and Gao et al. (2004) exploit syntactic or semantic parsers to capture long-range dependencies in language.”
A bidirectional target-filtering model of speech coarticulation and reduction: two-stage implementation for phonetic recognition (2006, unknown): “HE IMPORTANCE of incorporating structures of human speech and language into statistical models for technology applications has been well known, and active research in this direction has been pursued in recent years [2], [3], [4], [6], [10], [11], [19], [27], [31], [33].”
Structured speech modeling (2006, methodology): “Syntactic structure is exploited for characterizing the long-distance relationship among words in [5], [54].”
Talking to machines (statistically speaking) (2002, methodology): “In [29], Chelba and Mahajan describe how a structured language model [30] intialised on general treebank data can be trained for a speciﬁc SDS. Essentially this approach is similar to Wang and Acero in that a two level grammar is learnt.”
Continuous space language models (2007, methodology): “In the domain of language modeling for continuous speech recognition, a large body of interesting work has been carried out and many new approaches have been developed, for instance structured language models (Chelba and Jelinek, 2000) or maximum entropy language models (Rosenfeld, 1996).”
Random forests and the data sparseness problem in language modeling (2007, methodology): “This setup and the 50-best lists have been used in several previous studies (Chelba and Jelinek, 2000; Roark, 2001; Chelba and Xu, 2001; Xu et al., 2002).”
Reading Level Assessment Using Support Vector Machines and Statistical Language Models (2005, methodology): “Future work includes testing additional classi(cid:2)er features, e.g. parser likelihood scores and features obtained using a syntax-based language model such as Chelba and Jelinek (2000) or Roark (2001).”
Automatic speech recognition for under-resourced languages: A survey (2014, background): “There are some recent works that suggest taking into account syntactical information and long distance dependencies between words in sentences simultaneously with statistical language modeling, for instance, structured language models (Chelba and Jelinek, 2000) and some enhanced n-gram models (Kanej”
Modelling and Optimizing on Syntactic N-Grams for Statistical Machine Translation (2015, background): “, 2003), or structured language models (Chelba and Jelinek, 2000); previous efforts to include them in the translation process failed to improve translation performance (Och et al.”
Statistical language model adaptation: review and perspectives (2004, methodology): “One way to implement this strategy is to exploit the structured language model framework (Chelba and Jelinek, 2000; Jelinek and Chelba, 1999).”
A Neural Syntactic Language Model (2005, methodology): “Even though it maximizes an approximate objective function, it still can be proved that the “N-best” EM algorithm does converge to a local maximum (see Chelba & Jelinek, 2000 ) for discussion and comments).”
Exploiting prosodic breaks in language modeling with random forests (2008, background): “, [22, 23]) or tried to find the optimal back-off path or combination of paths [24, 25].”
A Challenge Set for Advancing Language Modeling (2012, background): “…of N-gram modeling techniques, there has been an important stream of research focused on the incorporation of syntactic and semantic information (Chelba and Jelinek, 1998; Chelba and Jelinek, 2000; Rosenfeld et al., 2001; Yamada and Knight, 2001; Khudanpur and Wu, 2000; Wu and Khudanpur, 1999).”
Recurrent neural network based language model personalization by social network crowdsourcing (2013, methodology): “In traditional N-gram models, the two corpora are interpolated; various methods are used to estimate their interpolation weights [35, 36, 37, 38].”
A Bayesian Model for Generative Transition-based Dependency Parsing (2015, background): “Syntactic language models have also been shown to improve performance in speech recognition and machine translation (Chelba and Jelinek, 2000; Charniak et al., 2003).”
Generative Incremental Dependency Parsing with Neural Networks (2015, methodology): “Syntactic language models based on PCFGs (Roark, 2001; Charniak, 2001) and incremental parsing (Chelba and Jelinek, 2000; Emami and Jelinek, 2005) have been proposed for speech recognition and machine translation.”