We address two challenges in topic models: (1) Context information around words helps in determining their actual meaning, e.g., "networks" used in the contexts "artificial neural networks" vs. "biological neuron networks". Generative topic models infer topic-word distributions, taking no or only little context into account. Here, we extend a neural autoregressive topic model to exploit the full context information around words in a document in a language modeling fashion. The proposed model is named as iDocNADE. (2) Due to the small number of word occurrences (i.e., lack of context) in short text and data sparsity in a corpus of few documents, the application of topic models is challenging on such texts. Therefore, we propose a simple and efficient way of incorporating external knowledge into neural autoregressive topic models: we use embeddings as a distributional prior. The proposed variants are named as DocNADEe and iDocNADEe. We present novel neural autoregressive topic model variants that consistently outperform state-of-the-art generative topic models in terms of generalization, interpretability (topic coherence) and applicability (retrieval and classification) over 7 long-text and 8 short-text datasets from diverse domains.
Conventional topic models are ineffective for topic extraction from microblog messages, because the data sparseness exhibited in short messages lacking structure and contexts results in poor message-level word co-occurrence patterns. To address this issue, we organize microblog messages as conversation trees based on their reposting and replying relations, and propose an unsupervised model that jointly learns word distributions to represent: 1) different roles of conversational discourse, 2) various latent topics in reflecting content information. By explicitly distinguishing the probabilities of messages with varying discourse roles in containing topical words, our model is able to discover clusters of discourse words that are indicative of topical content. In an automatic evaluation on large-scale microblog corpora, our joint model yields topics with better coherence scores than competitive topic models from previous studies. Qualitative analysis on model outputs indicates that our model induces meaningful representations for both discourse and topics. We further present an empirical study on microblog summarization based on the outputs of our joint model. The results show that the jointly modeled discourse and topic representations can effectively indicate summary-worthy content in microblog conversations.
Topic models are evaluated based on their ability to describe documents well (i.e. low perplexity) and to produce topics that carry coherent semantic meaning. In topic modeling so far, perplexity is a direct optimization target. However, topic coherence, owing to its challenging computation, is not optimized for and is only evaluated after training. In this work, under a neural variational inference framework, we propose methods to incorporate a topic coherence objective into the training process. We demonstrate that such a coherence-aware topic model exhibits a similar level of perplexity as baseline models but achieves substantially higher topic coherence.
A hallmark of variational autoencoders (VAEs) for text processing is their combination of powerful encoder-decoder models, such as LSTMs, with simple latent distributions, typically multivariate Gaussians. These models pose a difficult optimization problem: there is an especially bad local optimum where the variational posterior always equals the prior and the model does not use the latent variable at all, a kind of "collapse" which is encouraged by the KL divergence term of the objective. In this work, we experiment with another choice of latent distribution, namely the von Mises-Fisher (vMF) distribution, which places mass on the surface of the unit hypersphere. With this choice of prior and posterior, the KL divergence term now only depends on the variance of the vMF distribution, giving us the ability to treat it as a fixed hyperparameter. We show that doing so not only averts the KL collapse, but consistently gives better likelihoods than Gaussians across a range of modeling conditions, including recurrent language modeling and bag-of-words document modeling. An analysis of the properties of our vMF representations shows that they learn richer and more nuanced structures in their latent representations than their Gaussian counterparts.
We introduce supervised latent Dirichlet allocation (sLDA), a statistical model of labelled documents. The model accommodates a variety of response types. We derive an approximate maximum-likelihood procedure for parameter estimation, which relies on variational methods to handle intractable posterior expectations. Prediction problems motivate this research: we use the fitted model to predict response values for new documents. We test sLDA on two real-world problems: movie ratings predicted from reviews, and the political tone of amendments in the U.S. Senate based on the amendment text. We illustrate the benefits of sLDA versus modern regularized regression, as well as versus an unsupervised LDA analysis followed by a separate regression.
The growing need to analyze large collections of documents has led to great developments in topic modeling. Since documents are frequently associated with other related variables, such as labels or ratings, much interest has been placed on supervised topic models. However, the nature of most annotation tasks, prone to ambiguity and noise, often with high volumes of documents, deem learning under a single-annotator assumption unrealistic or unpractical for most real-world applications. In this article, we propose two supervised topic models, one for classification and another for regression problems, which account for the heterogeneity and biases among different annotators that are encountered in practice when learning from crowds. We develop an efficient stochastic variational inference algorithm that is able to scale to very large datasets, and we empirically demonstrate the advantages of the proposed model over state-of-the-art approaches.
In this technical report, we present jLDADMM---an easy-to-use Java toolkit for conventional topic models. jLDADMM is released to provide alternatives for topic modeling on normal or short texts. It provides implementations of the Latent Dirichlet Allocation topic model and the one-topic-per-document Dirichlet Multinomial Mixture model (i.e. mixture of unigrams), using collapsed Gibbs sampling. In addition, jLDADMM supplies a document clustering evaluation to compare topic models. jLDADMM is open-source and available to download at: https://github.com/datquocnguyen/jLDADMM
Context information around words helps in determining their actual meaning, for example "networks" used in contexts of artificial neural networks or biological neuron networks. Generative topic models infer topic-word distributions, taking no or only little context into account. Here, we extend a neural autoregressive topic model to exploit the full context information around words in a document in a language modeling fashion. This results in an improved performance in terms of generalization, interpretability and applicability. We apply our modeling approach to seven data sets from various domains and demonstrate that our approach consistently outperforms stateof-the-art generative topic models. With the learned representations, we show on an average a gain of 9.6% (0.57 Vs 0.52) in precision at retrieval fraction 0.02 and 7.2% (0.582 Vs 0.543) in F1 for text categorization.
In the last decade, a variety of topic models have been proposed for text engineering. However, except Probabilistic Latent Semantic Analysis (PLSA) and Latent Dirichlet Allocation (LDA), most of existing topic models are seldom applied or considered in industrial scenarios. This phenomenon is caused by the fact that there are very few convenient tools to support these topic models so far. Intimidated by the demanding expertise and labor of designing and implementing parameter inference algorithms, software engineers are prone to simply resort to PLSA/LDA, without considering whether it is proper for their problem at hand or not. In this paper, we propose a configurable topic modeling framework named Familia, in order to bridge the huge gap between academic research fruits and current industrial practice. Familia supports an important line of topic models that are widely applicable in text engineering scenarios. In order to relieve burdens of software engineers without knowledge of Bayesian networks, Familia is able to conduct automatic parameter inference for a variety of topic models. Simply through changing the data organization of Familia, software engineers are able to easily explore a broad spectrum of existing topic models or even design their own topic models, and find the one that best suits the problem at hand. With its superior extendability, Familia has a novel sampling mechanism that strikes balance between effectiveness and efficiency of parameter inference. Furthermore, Familia is essentially a big topic modeling framework that supports parallel parameter inference and distributed parameter storage. The utilities and necessity of Familia are demonstrated in real-life industrial applications. Familia would significantly enlarge software engineers' arsenal of topic models and pave the way for utilizing highly customized topic models in real-life problems.
Collaborative filtering based recommender systems have proven to be extremely successful in settings where user preference data on items is abundant. However, collaborative filtering algorithms are hindered by their weakness against the item cold-start problem and general lack of interpretability. Ontology-based recommender systems exploit hierarchical organizations of users and items to enhance browsing, recommendation, and profile construction. While ontology-based approaches address the shortcomings of their collaborative filtering counterparts, ontological organizations of items can be difficult to obtain for items that mostly belong to the same category (e.g., television series episodes). In this paper, we present an ontology-based recommender system that integrates the knowledge represented in a large ontology of literary themes to produce fiction content recommendations. The main novelty of this work is an ontology-based method for computing similarities between items and its integration with the classical Item-KNN (K-nearest neighbors) algorithm. As a study case, we evaluated the proposed method against other approaches by performing the classical rating prediction task on a collection of Star Trek television series episodes in an item cold-start scenario. This transverse evaluation provides insights into the utility of different information resources and methods for the initial stages of recommender system development. We found our proposed method to be a convenient alternative to collaborative filtering approaches for collections of mostly similar items, particularly when other content-based approaches are not applicable or otherwise unavailable. Aside from the new methods, this paper contributes a testbed for future research and an online framework to collaboratively extend the ontology of literary themes to cover other narrative content.
The task of discovering topics in text corpora has been dominated by Latent Dirichlet Allocation and other Topic Models for over a decade. In order to apply these approaches to massive text corpora, the vocabulary needs to be reduced considerably and large computer clusters and/or GPUs are typically required. Moreover, the number of topics must be provided beforehand but this depends on the corpus characteristics and it is often difficult to estimate, especially for massive text corpora. Unfortunately, both topic quality and time complexity are sensitive to this choice. This paper describes an alternative approach to discover topics based on Min-Hashing, which can handle massive text corpora and large vocabularies using modest computer hardware and does not require to fix the number of topics in advance. The basic idea is to generate multiple random partitions of the corpus vocabulary to find sets of highly co-occurring words, which are then clustered to produce the final topics. In contrast to probabilistic topic models where topics are distributions over the complete vocabulary, the topics discovered by the proposed approach are sets of highly co-occurring words. Interestingly, these topics underlie various thematics with different levels of granularity. An extensive qualitative and quantitative evaluation using the 20 Newsgroups (18K), Reuters (800K), Spanish Wikipedia (1M), and English Wikipedia (5M) corpora shows that the proposed approach is able to consistently discover meaningful and coherent topics. Remarkably, the time complexity of the proposed approach is linear with respect to corpus and vocabulary size; a non-parallel implementation was able to discover topics from the entire English edition of Wikipedia with over 5 million documents and 1 million words in less than 7 hours.
Classic Topic Models are built under the Bag Of Words assumption, in which word position is ignored for simplicity. Besides, symmetric priors are typically used in most applications. In order to easily learn topics with different properties among the same corpus, we propose a new line of work in which the paragraph structure is exploited. Our proposal is based on the following assumption: in many text document corpora there are formal constraints shared across all the collection, e.g. sections. When this assumption is satisfied, some paragraphs may be related to general concepts shared by all documents in the corpus, while others would contain the genuine description of documents. Assuming each paragraph can be semantically more general, specific, or hybrid, we look for ways to measure this, transferring this distinction to topics and being able to learn what we call specific and general topics. Experiments show that this is a proper methodology to highlight certain paragraphs in structured documents at the same time we learn interesting and more diverse topics.
With the considerable development of customer-to-customer (C2C) e-commerce in the recent years, there is a big demand for an effective recommendation system that suggests suitable websites for users to sell their items with some specified needs. Nonetheless, e-commerce recommendation systems are mostly designed for business-to-customer (B2C) websites, where the systems offer the consumers the products that they might like to buy. Almost none of the related research works focus on choosing selling sites for target items. In this paper, we introduce an approach that recommends the selling websites based upon the item's description, category, and desired selling price. This approach employs NoSQL data-based machine learning techniques for building and training topic models and classification models. The trained models can then be used to rank the websites dynamically with respect to the user needs. The experimental results with real-world datasets from Vietnam C2C websites will demonstrate the effectiveness of our proposed method.
People belong to multiple communities, words belong to multiple topics, and books cover multiple genres; overlapping clusters are commonplace. Many existing overlapping clustering methods model each person (or word, or book) as a non-negative weighted combination of "exemplars" who belong solely to one community, with some small noise. Geometrically, each person is a point on a cone whose corners are these exemplars. This basic form encompasses the widely used Mixed Membership Stochastic Blockmodel of networks (Airoldi et al., 2008) and its degree-corrected variants (Jin et al., 2017), as well as topic models such as LDA (Blei et al., 2003). We show that a simple one-class SVM yields provably consistent parameter inference for all such models, and scales to large datasets. Experimental results on several simulated and real datasets show our algorithm (called SVM-cone) is both accurate and scalable.
Stochastic gradient Markov chain Monte Carlo (SGMCMC) has become a popular method for scalable Bayesian inference. These methods are based on sampling a discrete-time approximation to a continuous time process, such as the Langevin diffusion. When applied to distributions defined on a constrained space the time-discretization error can dominate when we are near the boundary of the space. We demonstrate that because of this, current SGMCMC methods for the simplex struggle with sparse simplex spaces; when many of the components are close to zero. Unfortunately, many popular large-scale Bayesian models, such as network or topic models, require inference on sparse simplex spaces. To avoid the biases caused by this discretization error, we propose the stochastic Cox-Ingersoll-Ross process (SCIR), which removes all discretization error and we prove that samples from the SCIR process are asymptotically unbiased. We discuss how this idea can be extended to target other constrained spaces. Use of the SCIR process within a SGMCMC algorithm is shown to give substantially better performance for a topic model and a Dirichlet process mixture model than existing SGMCMC approaches.
Inference is an integral part of probabilistic topic models, but is often non-trivial to derive an efficient algorithm for a specific model. It is even much more challenging when we want to find a fast inference algorithm which always yields sparse latent representations of documents. In this article, we introduce a simple framework for inference in probabilistic topic models, denoted by FW. This framework is general and flexible enough to be easily adapted to mixture models. It has a linear convergence rate, offers an easy way to incorporate prior knowledge, and provides us an easy way to directly trade off sparsity against quality and time. We demonstrate the goodness and flexibility of FW over existing inference methods by a number of tasks. Finally, we show how inference in topic models with nonconjugate priors can be done efficiently.
In this paper we present a modification to a latent topic model, which makes the model exploit supervision to produce a factorized representation of the observed data. The structured parameterization separately encodes variance that is shared between classes from variance that is private to each class by the introduction of a new prior over the topic space. The approach allows for a more eff{}icient inference and provides an intuitive interpretation of the data in terms of an informative signal together with structured noise. The factorized representation is shown to enhance inference performance for image, text, and video classification.
We develop stochastic variational inference, a scalable algorithm for approximating posterior distributions. We develop this technique for a large class of probabilistic models and we demonstrate it with two probabilistic topic models, latent Dirichlet allocation and the hierarchical Dirichlet process topic model. Using stochastic variational inference, we analyze several large collections of documents: 300K articles from Nature, 1.8M articles from The New York Times, and 3.8M articles from Wikipedia. Stochastic inference can easily handle data sets of this size and outperforms traditional variational inference, which can only handle a smaller subset. (We also show that the Bayesian nonparametric topic model outperforms its parametric counterpart.) Stochastic variational inference lets us apply complex Bayesian models to massive data sets.
We give a robust version of the celebrated result of Kruskal on the uniqueness of tensor decompositions: we prove that given a tensor whose decomposition satisfies a robust form of Kruskal's rank condition, it is possible to approximately recover the decomposition if the tensor is known up to a sufficiently small (inverse polynomial) error. Kruskal's theorem has found many applications in proving the identifiability of parameters for various latent variable models and mixture models such as Hidden Markov models, topic models etc. Our robust version immediately implies identifiability using only polynomially many samples in many of these settings. This polynomial identifiability is an essential first step towards efficient learning algorithms for these models. Recently, algorithms based on tensor decompositions have been used to estimate the parameters of various hidden variable models efficiently in special cases as long as they satisfy certain "non-degeneracy" properties. Our methods give a way to go beyond this non-degeneracy barrier, and establish polynomial identifiability of the parameters under much milder conditions. Given the importance of Kruskal's theorem in the tensor literature, we expect that this robust version will have several applications beyond the settings we explore in this work.
In the internet era there has been an explosion in the amount of digital text information available, leading to difficulties of scale for traditional inference algorithms for topic models. Recent advances in stochastic variational inference algorithms for latent Dirichlet allocation (LDA) have made it feasible to learn topic models on large-scale corpora, but these methods do not currently take full advantage of the collapsed representation of the model. We propose a stochastic algorithm for collapsed variational Bayesian inference for LDA, which is simpler and more efficient than the state of the art method. We show connections between collapsed variational Bayesian inference and MAP estimation for LDA, and leverage these connections to prove convergence properties of the proposed algorithm. In experiments on large-scale text corpora, the algorithm was found to converge faster and often to a better solution than the previous method. Human-subject experiments also demonstrated that the method can learn coherent topics in seconds on small corpora, facilitating the use of topic models in interactive document analysis software.