Topic Models
210 papers with code • 6 benchmarks • 12 datasets
A topic model is a type of statistical model for discovering the abstract "topics" that occur in a collection of documents. Topic modeling is a frequently used text-mining tool for the discovery of hidden semantic structures in a text body.
Benchmarks
These leaderboards are used to track progress in Topic Models
| Trend | Dataset | Best Model | Paper | Code | Compare |
|---|---|---|---|---|---|
|
|||||
|
|||||
|
|||||
|
|||||
|
|||||
|
Libraries
Use these libraries to find Topic Models models and implementationsDatasets
AG News
Reddit
New York Times Annotated Corpus
Arxiv HEP-TH citation graph
OpoSum
Reddit Ideology Database
Latest papers
Concept Induction: Analyzing Unstructured Text with High-Level Concepts Using LLooM
Data analysts have long sought to turn unstructured text data into meaningful concepts.
GINopic: Topic Modeling with Graph Isomorphism Network
Topic modeling is a widely used approach for analyzing and exploring large document collections.
Enhanced Short Text Modeling: Leveraging Large Language Models for Topic Refinement
Crafting effective topic models for brief texts, like tweets and news headlines, is essential for capturing the swift shifts in social dynamics.
Neural Multimodal Topic Modeling: A Comprehensive Evaluation
This paper presents the first systematic and comprehensive evaluation of multimodal topic modeling of documents containing both text and images.
Automating the Information Extraction from Semi-Structured Interview Transcripts
likich/tm_graph
•
•
This paper explores the development and application of an automated system designed to extract information from semi-structured interview transcripts.
Membership Inference Attacks and Privacy in Topic Modeling
Recent research shows that large language models are susceptible to privacy attacks that infer aspects of the training data.
Network-based Topic Structure Visualization
In the real world, many topics are inter-correlated, making it challenging to investigate their structure and relationships.
Improving the TENOR of Labeling: Re-evaluating Topic Models for Content Analysis
Topic models are a popular tool for understanding text collections, but their evaluation has been a point of contention.
A Survey on Neural Topic Models: Methods, Applications, and Challenges
bobxwu/topmost
•
•
In this paper, we present a comprehensive survey on neural topic models concerning methods, applications, and challenges.
Revisiting Topic-Guided Language Models
carolinazheng/revisiting-tglms
•
•
A recent line of work in natural language processing has aimed to combine language models and topic models.
