Integrating Topics and Syntax

By T. Griffiths, Mark Steyvers, David M. Blei, Josh Tenenbaum. Neural Information Processing Systems, 2004

628 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/0ecc5ffeae38689dd2fe6ed4c32a6745744d7641

How researchers see this paper

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

Citations in the corpus

Stochastic Variational Inference for Bayesian Time Series Models (2014, background): “, 2013), brain-machine interfaces (Hudson, 2008), and natural language and text (Griffiths et al., 2004; Liang et al., 2007).”
Mining Text Data (2012, methodology): “It has also been shown both theoretically and experimentally [33, 93] that the matrix-factorization technique is equivalent to another graph-structure based document clustering technique known”
Sparse Additive Generative Models of Text (2011, background): “• Inference cost There is increasing interest in modeling text with multiple generative facets, such as syntax (Griffiths et al., 2005), sentiment (Mei et al.”
Bayesian models of cognition (2008, methodology): “…acquisition (Chater & Manning, 2006; Xu & Tenenbaum, in press), symbolic reasoning (Oaksford & Chater, 2001), causal learning and inference (Steyvers, Tenenbaum, Wagenmakers, & Blum, 2003; Griﬃths & Tenenbaum, 2005, 2007a), and social cognition (Baker, Tenenbaum, & Saxe, 2007), among other topics.”
Syntactic Topic Models (2008, background): “Griffiths et al (Griffiths et al. 2005) observed that nouns, more than other parts of speech, tend to specialize into distinct topics, and this is also evident here.”
Generating More Interesting Responses in Neural Conversation Models with Distributional Constraints (2018, background): “We propose two constraints that help generate more content rich responses that are based on a model of syntax and topics (Griffiths et al., 2005) and semantic similarity (Arora et al.”
Correlated Topic Models (2005, methodology): “Recent work has used LDA in more complicated document models [9, 11, 7], and in a variety of settings such as image processing [12], collaborative filtering [8], and the modeling of sequential data and user profiles [6].”
Efficient indexing for large scale visual search (2009, methodology): “Gibbs sampling is a popular method to sample complex posterior [8].”
Evaluating probabilities under high-dimensional latent variable models (2008, methodology): “Despite these problems the estimator has received significant attention in statistics, and has been used for evaluating latent variable models in recent machine learning literature [5, 6].”
Semi-Supervised Sequence Modeling with Syntactic Topic Models (2005, background): “HMM-LDA (Griffiths et al. 2005) is a probabilistic generative model that distinguishes between function words and content words.”
Topic and Role Discovery in Social Networks (2005, background): “New topic models have been actively studied in recent years for many dierent tasks, including joint modeling of words and research paper citations (Erosheva, Fienberg, & Laerty, 2004), capturing correlations among topics (Blei & Laerty, 2006a; Li & McCallum, 2006), taking advantage of both topical a”
Pachinko allocation: DAG-structured mixture models of topic correlations (2006, methodology): “Another variation of LDA is the HMMLDA model (Griffiths et al., 2005), which combines a hidden Markov model (HMM) with LDA to extract word clusters from sequential data.”
Rethinking LDA: Why Priors Matter (2009, methodology): “The robustness of AS to stop words has implications for HMM-LDA [8] which models stop words using a hidden Markov model and “content” words using LDA, at considerable computational cost.”
Topic modeling: beyond bag-of-words (2006, result): “This effect is similar, though less pronounced, to that achieved by Griffiths et al.’s composite model (2004), in which function words are handled by a hidden Markov model, while content words are handled by latent Dirichlet allocation.”
Applications of Topics Models to Analysis of Disaster-Related Twitter Data (2009, background): “Topics models are usually "bag-of-words" models, meaning that they do not rely on syntactic structure or word order in language (though can be adapted for doing so [2]) Thus, they are likely to be better suited to handle esoteric language and irregular grammar of typical Twitter messages.”
Topic Modeling in Embedding Spaces (2019, result): “to cope with stop words (Griffiths et al., 2004; Chemudugunta et al., 2006;Wallach et al., 2009a); our goal here is not to establish comparisons with these methods but to show the performance of the ETM-PWE in the presence of stop words.”