The Author-Topic Model for Authors and Documents
By Michal Rosen-Zvi, T. Griffiths, Mark Steyvers, Padhraic Smyth. Conference on Uncertainty in Artificial Intelligence, 2004
1,706 citations recorded by the source corpus.
We introduce the author-topic model, a generative model for documents that extends Latent Dirichlet Allocation (LDA; Blei, Ng, & Jordan, 2003) to include authorship information. Each author is associated with a multinomial distribution over topics and each topic is associated with a multinomial distribution over words. A document with multiple authors is modeled as a distribution over topics that is a mixture of the distributions associated with the authors. We apply the model to a collection of 1,700 NIPS conference papers and 160,000 CiteSeer abstracts. Exact inference is intractable for these datasets and we use Gibbs sampling to estimate the topic and author distributions. We compare the performance with two other generative models for documents, which are special cases of the author-topic model: LDA (a topic model) and a simple author model in which each author is associated with a distribution over words rather than a distribution over topics. We show topics recovered by the author-topic model, and demonstrate applications to computing similarity between authors and entropy of author output.
Subfields: Artificial Intelligence, General Social Sciences
View Paper: https://www.semanticscholar.org/paper/6436dce0e39f15a1ca9269e6ca813dfebb0af3a2
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
- Yoshua Bengio
- Jianfeng Gao
- Mohit Bansal
- Ruslan Salakhutdinov
- Percy Liang
- Quoc V. Le
- Kyunghyun Cho
- Sergey Levine
In General Social Sciences
- David M. Blei
- David Mimno
- Chengxiang Zhai
- Charu C. Aggarwal
- Chong Wang
- Andrew McCallum
- Jordan L. Boyd-Graber
- Jon D. McAuliffe
Citations in the corpus
- Statistical topic models for multi-label document classification (2011, methodology): “ correspondence with labels and then restricting the sampling of topics for each document to the set of labels that were assigned to the document, in a manner similar to the Author-Model described by Rosen-Zvi et al. (2004) (where the set of authors for each document in the Author Model is now repla”
- An overview of topic modeling and its current applications in bioinformatics (2016, background): “In the author-topic model (Rosen-Zvi et al. 2004), the generative process is as follows: choose an author at random; generate a word based on the topic probability distribution of this author; repeat the above steps until the document generation is finished.”
- Block-LDA: Jointly Modeling Entity-Annotated Text and Entity-Entity Links (2014, methodology): “The Author-Topic model (AT) [12] addresses the task of modeling corpora annotated with the ids of people who authored the documents.”
- Expertise modeling for matching papers with reviewers (2007, methodology): “These methods include language models with Dirichlet smoothing (Ponte & Croft, 1998; Zhai & Lafferty, 2001), the Author-Topic model (Rosen-Zvi et al., 2004), and a novel topic model, Author-Persona-Topic (APT).”
- Short Text Topic Modeling Techniques, Applications, and Performance: A Survey (2019, methodology): “The second step estimates draws a pair of pseudo-document label l d,w and topic label z d,w jointly for word w in document d , which is similar with standard topic modeling (author-topic modeling) [42].”
- Exponential Family Harmoniums with an Application to Information Retrieval (2004, methodology): “The EFH and LSI models were trained as in the previous experiment while the training and testing details 7 for LDA can be found in [9].”
- Topic Models Conditioned on Arbitrary Features with Dirichlet-multinomial Regression (2008, methodology): “One example of this type of model is the author-topic model of Rosen-Zvi, Griffiths, Steyvers and Smyth [13].”
- A model of shared grasp affordances from demonstration (2007, methodology): “The model of Sudderth is an adaptation of statistical topic models such as the author-topic model [5] and Latent Dirichlet Allocation (LDA) [2].”
- Learning author-topic models from text corpora (2010, methodology): “[2004] and Rosen-Zvi et al. [2004], by providing a systematic comparison of the author-topic model to other existing models, showing how the authortopic model can be applied as an extension of the LDA model with the use of fictitious authors and illustrating a number of applications of the model.”
- A correlated topic model of Science (2007, methodology): “Recent work has used LDA as a building block in more sophisticated topic models, such as author-document models [19, 24], abstract-reference models [12] syntaxsemantics models [16] and image-caption models [6].”
- A Survey on Context Learning (2017, background): “[41] introduces the author-topic model and discusses how to utilize the author context information when analyzing document collections.”
- Supervised Topic Models (2007, methodology): “Third, the technique we have used to incorporate a response can be applied in existing variants of LDA, such as author-topic models (Rosen-Zvi et al., 2004), population-genetics models (Pritchard et al., 2000), and survey-data models (Erosheva, 2002).”
- Mining Text Data (2012, methodology): “The detailed representation would likely enable the development of better summarizers which conveys the similarities and differences among the different documents that make up the input for multi-document summarization [55, 24, 54].”
- Learning hierarchical models of scenes, objects, and parts (2005, methodology): “To further investigate these shared parts, we used the symmetrized KL divergence [15] to compute a distance between all pairs of object–specific part distributions:”
- DiscLDA: Discriminative Learning for Dimensionality Reduction and Classification (2008, methodology): “By including a Dirichlet prior on theT parameters, the DiscLDA model can be related to the authortopic model [9], if we restrict to the special case in which th ere is only one author per document.”