Optimizing Semantic Coherence in Topic Models

By David Mimno, Hanna M. Wallach, E. Talley, Miriam Leenders, Andrew McCallum. Conference on Empirical Methods in Natural Language Processing, 2011

1,965 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/ef2d64e448ee5ed2dc26179c8570803ded123a5e

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

A Survey of App Store Analysis for Software Engineering (2017, methodology): “IDF, and not LDA or similar topic models such as the weighted topic model [162].”
Discovery of Treatments from Text Corpora (2016, methodology): “To create a candidate set of models, we use a quantitative measure that balances coherence and exclusivity (Mimno et al., 2011; Roberts et al., 2014).”
Short Text Topic Modeling Techniques, Applications, and Performance: A Survey (2019, background): “A lot of metrics have been proposed for measuring the coherence of topics in texts [59], [60].”
Learning from Your Network of Friends: A Trajectory Representation Learning Model Based on Online Social Ties (2016, methodology): “Additionally, our proposed model is closely related to two topic models, LDA with optimized parameters [8] and coherent LDA [9].”
Topic Modeling using Topics from Many Domains, Lifelong Learning and Big Data (2014, methodology): “measure (Mimno et al., 2011) is proposed as a better alternative for assessing topic quality.”
A biterm topic model for short texts (2013, methodology): “Among them two works close to us are the recently proposed regularized topic model [22] and the generalized P´olya model [21], which also employ word co-occurrence statistics to enhance topic learning.”
Computer-Assisted Text Analysis for Comparative Politics (2015, methodology): “Every topic has some level of correlation with every other topic, but intuitively Specifically we use the semantic coherence metric developed by (Mimno et al., 2011).”
Topic Compositional Neural Language Model (2017, methodology): “We evaluate the topic model by inspecting the coherence of inferred topics (Chang et al., 2009; New-man et al., 2010; Mimno et al., 2011).”
Graph-Sparse LDA: A Topic Model with Structured Sparsity (2014, background): “[22] and [23] developed me asures of topic coherence that strongly correlated with human annotations of topic quality.”
Synthetic Lies: Understanding AI-Generated Misinformation and Evaluating Algorithmic and Human Solutions (2023, background): “a topic tend to co-occur, with evidence showing it correlates with expert opinions of topic quality [73].”
Exploring the Space of Topic Coherence Measures (2015, methodology): “, humans were asked to classify word sets using different given error types [12].”
LDAvis: A method for visualizing and interpreting topics (2014, methodology): “AlSumait et al. (2009), Mimno et al. (2011), and Chuang et al. (2013b) develop quantitative methods for measuring the interpretability of top-64 ics based on experiments with data sets that come with some notion of topical ground truth, such as document metadata or expert-created topic labels.”
Interpretable Machine Learning : Lessons from Topic Modeling (2016, background): “Metrics to measure the coherence of topics were introduced by [17] and [16], and later improved upon and generalized by others [13] [21].”
An Unsupervised Neural Attention Model for Aspect Extraction (2017, methodology): “In order to objectively measure the quality of aspects, we use coherence score as a metric which has been shown to correlate well with human judgment (Mimno et al., 2011).”
Applying LDA Topic Modeling in Communication Research: Toward a Valid and Reliable Methodology (2018, methodology): “Studies have shown that coherence measured with respect to data that is external (Newman, Lau, Grieser, & Baldwin, 2010) or internal to the corpus (Mimno et al., 2011) correlates with human judgment on topic interpretability.”
A Practical Algorithm for Topic Modeling with Provable Guarantees (2012, methodology): “Coherence is based on two functions, D ( w ) and D ( w 1 , w 2 ), which are number of documents with at least one instance of w , and of w 1 and w 2 , respectively [Mimno et al., 2011].”
An analysis of the coherence of descriptors in topic modeling (2015, methodology): “…Correlated Topic Model (CTM) of Blei and Lafferty (2006), which attempts to directly model correlation between the latent topics themselves, and the Pólya Urn method proposed by Mimno et al. (2011), which extended Gibbs sampling to incorporate information used in the corresponding coherence metric.”
Mining Aspect-Specific Opinion using a Holistic Lifelong Topic Model (2016, methodology): “The modeling of traditional topic models, such as LDA, is equivalent to the simple Pólya urn model [31].”