Millions of online discussions are generated everyday on social media platforms. Topic modelling is an efficient way of better understanding large text datasets at scale. Conventional topic models have had limited success in online discussions, and to overcome their limitations, we use the discussion thread tree structure and propose a "popularity" metric to quantify the number of replies to a comment to extend the frequency of word occurrences, and the "transitivity" concept to characterize topic dependency among nodes in a nested discussion thread. We build a Conversational Structure Aware Topic Model (CSATM) based on popularity and transitivity to infer topics and their assignments to comments. Experiments on real forum datasets are used to demonstrate improved performance for topic extraction with six different measurements of coherence and impressive accuracy for topic assignments.
Topic models have become popular tools for dimension reduction and exploratory analysis of text data which consists in observed frequencies of a vocabulary of $p$ words in $n$ documents, stored in a $p\times n$ matrix. The main premise is that the mean of this data matrix can be factorized into a product of two non-negative matrices: a $p\times K$ word-topic matrix $A$ and a $K\times n$ topic-document matrix $W$. This paper studies the estimation of $A$ that is possibly element-wise sparse, and the number of topics $K$ is unknown. In this under-explored context, we derive a new minimax lower bound for the estimation of such $A$ and propose a new computationally efficient algorithm for its recovery. We derive a finite sample upper bound for our estimator, and show that it matches the minimax lower bound in many scenarios. Our estimate adapts to the unknown sparsity of $A$ and our analysis is valid for any finite $n$, $p$, $K$ and document lengths. Empirical results on both synthetic data and semi-synthetic data show that our proposed estimator is a strong competitor of the existing state-of-the-art algorithms for both non-sparse $A$ and sparse $A$, and has superior performance is many scenarios of interest.
Document networks are found in various collections of real-world data, such as citation networks, hyperlinked web pages, and online social networks. A large number of generative models have been proposed because they offer intuitive and useful pictures for analyzing document networks. Prominent examples are relational topic models, where documents are linked according to their topic similarities. However, existing generative models do not make full use of network structures because they are largely dependent on topic modeling of documents. In particular, centrality of graph nodes is missing in generative processes of previous models. In this paper, we propose a novel generative model for document networks by introducing random walkers on networks to integrate the node centrality into link generation processes. The developed method is evaluated in semi-supervised classification tasks with real-world citation networks. We show that the proposed model outperforms existing probabilistic approaches especially in detecting communities in connected networks.
Topic modeling has found wide application in many problems where latent structures of the data are crucial for typical inference tasks. When applying a topic model, a relatively standard pre-processing step is to first build a vocabulary of frequent words. Such a general pre-processing step is often independent of the topic modeling stage, and thus there is no guarantee that the pre-generated vocabulary can support the inference of some optimal (or even meaningful) topic models appropriate for a given task, especially for computer vision applications involving "visual words". In this paper, we propose a new approach to topic modeling, termed Vocabulary-Selection-Embedded Correspondence-LDA (VSEC-LDA), which learns the latent model while simultaneously selecting most relevant words. The selection of words is driven by an entropy-based metric that measures the relative contribution of the words to the underlying model, and is done dynamically while the model is learned. We present three variants of VSEC-LDA and evaluate the proposed approach with experiments on both synthetic and real databases from different applications. The results demonstrate the effectiveness of built-in vocabulary selection and its importance in improving the performance of topic modeling.
Current variational inference methods for hierarchical Bayesian nonparametric models can neither characterize the correlation structure among latent variables due to the mean-field setting, nor infer the true posterior dimension because of the universal truncation. To overcome these limitations, we propose the conditional and adaptively truncated variational inference method (CATVI) by maximizing the nonparametric evidence lower bound and integrating Monte Carlo into the variational inference framework. CATVI enjoys several advantages over traditional methods, including a smaller divergence between variational and true posteriors, reduced risk of underfitting or overfitting, and improved prediction accuracy. Empirical studies on three large datasets reveal that CATVI applied in Bayesian nonparametric topic models substantially outperforms competing models, providing lower perplexity and clearer topic-words clustering.
The tremendous increase in the amount of available research documents impels researchers to propose topic models to extract the latent semantic themes of a documents collection. However, how to extract the hidden topics of the documents collection has become a crucial task for many topic model applications. Moreover, conventional topic modeling approaches suffer from the scalability problem when the size of documents collection increases. In this paper, the Correlated Topic Model with variational Expectation-Maximization algorithm is implemented in MapReduce framework to solve the scalability problem. The proposed approach utilizes the dataset crawled from the public digital library. In addition, the full-texts of the crawled documents are analysed to enhance the accuracy of MapReduce CTM. The experiments are conducted to demonstrate the performance of the proposed algorithm. From the evaluation, the proposed approach has a comparable performance in terms of topic coherences with LDA implemented in MapReduce framework.
Cross-lingual annotations of legislative texts enable us to explore major themes covered in multilingual legal data and are a key facilitator of semantic similarity when searching for similar documents. Multilingual probabilistic topic models have recently emerged as a group of semi-supervised machine learning models that can be used to perform thematic explorations on collections of texts in multiple languages. However, these approaches require theme-aligned training data to create a language-independent space, which limits the amount of scenarios where this technique can be used. In this work, we provide an unsupervised document similarity algorithm based on hierarchies of multi-lingual concepts to describe topics across languages. The algorithm does not require parallel or comparable corpora, or any other type of translation resource. Experiments performed on the English, Spanish, French and Portuguese editions of JCR-Acquis corpora reveal promising results on classifying and sorting documents by similar content.
Twitter data has been shown broadly applicable for public health surveillance. Previous public health studies based on Twitter data have largely relied on keyword-matching or topic models for clustering relevant tweets. However, both methods suffer from the short-length of texts and unpredictable noise that naturally occurs in user-generated contexts. In response, we introduce a deep learning approach that uses hashtags as a form of supervision and learns tweet embeddings for extracting informative textual features. In this case study, we address the specific task of estimating state-level obesity from dietary-related textual features. Our approach yields an estimation that strongly correlates the textual features to government data and outperforms the keyword-matching baseline. The results also demonstrate the potential of discovering risk factors using the textual features. This method is general-purpose and can be applied to a wide range of Twitter-based public health studies.
This research proposes a new (old) metric for evaluating goodness of fit in topic models, the coefficient of determination, or $R^2$. Within the context of topic modeling, $R^2$ has the same interpretation that it does when used in a broader class of statistical models. Reporting $R^2$ with topic models addresses two current problems in topic modeling: a lack of standard cross-contextual evaluation metrics for topic modeling and ease of communication with lay audiences. The author proposes that $R^2$ should be reported as a standard metric when constructing topic models.
One of the challenges for text analysis in medical domains is analyzing large-scale medical documents. As a consequence, finding relevant documents has become more difficult. One of the popular methods to retrieve information based on discovering the themes in the documents is topic modeling. The themes in the documents help to retrieve documents on the same topic with and without a query. In this paper, we present a novel approach to topic modeling using fuzzy clustering. To evaluate our model, we experiment with two text datasets of medical documents. The evaluation metrics carried out through document classification and document modeling show that our model produces better performance than LDA, indicating that fuzzy set theory can improve the performance of topic models in medical domains.
In this paper we present a model for unsupervised topic discovery in texts corpora. The proposed model uses documents, words, and topics lookup table embedding as neural network model parameters to build probabilities of words given topics, and probabilities of topics given documents. These probabilities are used to recover by marginalization probabilities of words given documents. For very large corpora where the number of documents can be in the order of billions, using a neural auto-encoder based document embedding is more scalable then using a lookup table embedding as classically done. We thus extended the lookup based document embedding model to continuous auto-encoder based model. Our models are trained using probabilistic latent semantic analysis (PLSA) assumptions. We evaluated our models on six datasets with a rich variety of contents. Conducted experiments demonstrate that the proposed neural topic models are very effective in capturing relevant topics. Furthermore, considering perplexity metric, conducted evaluation benchmarks show that our topic models outperform latent Dirichlet allocation (LDA) model which is classically used to address topic discovery tasks.
Making sense of words often requires to simultaneously examine the surrounding context of a term as well as the global themes characterizing the overall corpus. Several topic models have already exploited word embeddings to recognize local context, however, it has been weakly combined with the global context during the topic inference. This paper proposes to extract topical phrases corroborating the word embedding information with the global context detected by Latent Semantic Analysis, and then combine them by means of the P\'{o}lya urn model. To highlight the effectiveness of this combined approach the model was assessed analyzing clinical reports, a challenging scenario characterized by technical jargon and a limited word statistics available. Results show it outperforms the state-of-the-art approaches in terms of both topic coherence and computational cost.
Electronic Health Record (EHR) data can be represented as discrete counts over a high dimensional set of possible procedures, diagnoses, and medications. Supervised topic models present an attractive option for incorporating EHR data as features into a prediction problem: given a patient's record, we estimate a set of latent factors that are predictive of the response variable. However, existing methods for supervised topic modeling struggle to balance prediction quality and coherence of the latent factors. We introduce a novel approach, the prediction-focused topic model, that uses the supervisory signal to retain only features that improve, or do not hinder, prediction performance. By removing features with irrelevant signal, the topic model is able to learn task-relevant, interpretable topics. We demonstrate on a EHR dataset and a movie review dataset that compared to existing approaches, prediction-focused topic models are able to learn much more coherent topics while maintaining competitive predictions.
We consider the problem of aggregating models learned from sequestered, possibly heterogeneous datasets. Exploiting tools from Bayesian nonparametrics, we develop a general meta-modeling framework that learns shared global latent structures by identifying correspondences among local model parameterizations. Our proposed framework is model-independent and is applicable to a wide range of model types. After verifying our approach on simulated data, we demonstrate its utility in aggregating Gaussian topic models, hierarchical Dirichlet process based hidden Markov models, and sparse Gaussian processes with applications spanning text summarization, motion capture analysis, and temperature forecasting.
Topic modeling is widely studied for the dimension reduction and analysis of documents. However, it is formulated as a difficult optimization problem. Current approximate solutions also suffer from inaccurate model- or data-assumptions. To deal with the above problems, we propose a polynomial-time deep topic model with no model and data assumptions. Specifically, we first apply multilayer bootstrap network (MBN), which is an unsupervised deep model, to reduce the dimension of documents, and then use the low-dimensional data representations or their clustering results as the target of supervised Lasso for topic word discovery. To our knowledge, this is the first time that MBN and Lasso are applied to unsupervised topic modeling. Experimental comparison results with five representative topic models on the 20-newsgroups and TDT2 corpora illustrate the effectiveness of the proposed algorithm.
Supervised topic models are often sought to balance prediction quality and interpretability. However, when models are (inevitably) misspecified, standard approaches rarely deliver on both. We introduce a novel approach, the prediction-focused topic model, that uses the supervisory signal to retain only vocabulary terms that improve, or at least do not hinder, prediction performance. By removing terms with irrelevant signal, the topic model is able to learn task-relevant, coherent topics. We demonstrate on several data sets that compared to existing approaches, prediction-focused topic models learn much more coherent topics while maintaining competitive predictions.
Topic models such as LDA, DocNADE, iDocNADEe have been popular in document analysis. However, the traditional topic models have several limitations including: (1) Bag-of-words (BoW) assumption, where they ignore word ordering, (2) Data sparsity, where the application of topic models is challenging due to limited word co-occurrences, leading to incoherent topics and (3) No Continuous Learning framework for topic learning in lifelong fashion, exploiting historical knowledge (or latent topics) and minimizing catastrophic forgetting. This thesis focuses on addressing the above challenges within neural topic modeling framework. We propose: (1) Contextualized topic model that combines a topic and a language model and introduces linguistic structures (such as word ordering, syntactic and semantic features, etc.) in topic modeling, (2) A novel lifelong learning mechanism into neural topic modeling framework to demonstrate continuous learning in sequential document collections and minimizing catastrophic forgetting. Additionally, we perform a selective data augmentation to alleviate the need for complete historical corpora during data hallucination or replay.
Finding the right reviewers to assess the quality of conference submissions is a time consuming process for conference organizers. Given the importance of this step, various automated reviewer-paper matching solutions have been proposed to alleviate the burden. Prior approaches, including bag-of-words models and probabilistic topic models have been inadequate to deal with the vocabulary mismatch and partial topic overlap between a paper submission and the reviewer's expertise. Our approach, the common topic model, jointly models the topics common to the submission and the reviewer's profile while relying on abstract topic vectors. Experiments and insightful evaluations on two datasets demonstrate that the proposed method achieves consistent improvements compared to available state-of-the-art implementations of paper-reviewer matching.
Statistical topic models are increasingly and popularly used by Digital Humanities scholars to perform distant reading tasks on literary data. It allows us to estimate what people talk about. Especially Latent Dirichlet Allocation (LDA) has shown its usefulness, as it is unsupervised, robust, easy to use, scalable, and it offers interpretable results. In a preliminary study, we apply LDA to a corpus of New High German poetry (textgrid, with 51k poems, 8m token), and use the distribution of topics over documents for a classification of poems into time periods and for authorship attribution.
Distributed word embeddings have yielded state-of-the-art performance in many NLP tasks, mainly due to their success in capturing useful semantic information. These representations assign only a single vector to each word whereas a large number of words are polysemous (i.e., have multiple meanings). In this work, we approach this critical problem in lexical semantics, namely that of representing various senses of polysemous words in vector spaces. We propose a topic modeling based skip-gram approach for learning multi-prototype word embeddings. We also introduce a method to prune the embeddings determined by the probabilistic representation of the word in each topic. We use our embeddings to show that they can capture the context and word similarity strongly and outperform various state-of-the-art implementations.