Dynamic topic models (DTMs) are very effective in discovering topics and capturing their evolution trends in time series data. To do posterior inference of DTMs, existing methods are all batch algorithms that scan the full dataset before each update of the model and make inexact variational approximations with mean-field assumptions. Due to a lack of a more scalable inference algorithm, despite the usefulness, DTMs have not captured large topic dynamics. This paper fills this research void, and presents a fast and parallelizable inference algorithm using Gibbs Sampling with Stochastic Gradient Langevin Dynamics that does not make any unwarranted assumptions. We also present a Metropolis-Hastings based $O(1)$ sampler for topic assignments for each word token. In a distributed environment, our algorithm requires very little communication between workers during sampling (almost embarrassingly parallel) and scales up to large-scale applications. We are able to learn the largest Dynamic Topic Model to our knowledge, and learned the dynamics of 1,000 topics from 2.6 million documents in less than half an hour, and our empirical results show that our algorithm is not only orders of magnitude faster than the baselines but also achieves lower perplexity.
Supervised topic models simultaneously model the latent topic structure of large collections of documents and a response variable associated with each document. Existing inference methods are based on variational approximation or Monte Carlo sampling, which often suffers from the local minimum defect. Spectral methods have been applied to learn unsupervised topic models, such as latent Dirichlet allocation (LDA), with provable guarantees. This paper investigates the possibility of applying spectral methods to recover the parameters of supervised LDA (sLDA). We first present a two-stage spectral method, which recovers the parameters of LDA followed by a power update method to recover the regression model parameters. Then, we further present a single-phase spectral algorithm to jointly recover the topic distribution matrix as well as the regression weights. Our spectral algorithms are provably correct and computationally efficient. We prove a sample complexity bound for each algorithm and subsequently derive a sufficient condition for the identifiability of sLDA. Thorough experiments on synthetic and real-world datasets verify the theory and demonstrate the practical effectiveness of the spectral algorithms. In fact, our results on a large-scale review rating dataset demonstrate that our single-phase spectral algorithm alone gets comparable or even better performance than state-of-the-art methods, while previous work on spectral methods has rarely reported such promising performance.
Hyper-parameters play a major role in the learning and inference process of latent Dirichlet allocation (LDA). In order to begin the LDA latent variables learning process, these hyper-parameters values need to be pre-determined. We propose an extension for LDA that we call 'Latent Dirichlet allocation Gibbs Newton' (LDA-GN), which places non-informative priors over these hyper-parameters and uses Gibbs sampling to learn appropriate values for them. At the heart of LDA-GN is our proposed 'Gibbs-Newton' algorithm, which is a new technique for learning the parameters of multivariate Polya distributions. We report Gibbs-Newton performance results compared with two prominent existing approaches to the latter task: Minka's fixed-point iteration method and the Moments method. We then evaluate LDA-GN in two ways: (i) by comparing it with standard LDA in terms of the ability of the resulting topic models to generalize to unseen documents; (ii) by comparing it with standard LDA in its performance on a binary classification task.
Traditional topic models do not account for semantic regularities in language. Recent distributional representations of words exhibit semantic consistency over directional metrics such as cosine similarity. However, neither categorical nor Gaussian observational distributions used in existing topic models are appropriate to leverage such correlations. In this paper, we propose to use the von Mises-Fisher distribution to model the density of words over a unit sphere. Such a representation is well-suited for directional data. We use a Hierarchical Dirichlet Process for our base topic model and propose an efficient inference algorithm based on Stochastic Variational Inference. This model enables us to naturally exploit the semantic structures of word embeddings while flexibly discovering the number of topics. Experiments demonstrate that our method outperforms competitive approaches in terms of topic coherence on two different text corpora while offering efficient inference.
Topic models (e.g., pLSA, LDA, SLDA) have been widely used for segmenting imagery. These models are confined to crisp segmentation. Yet, there are many images in which some regions cannot be assigned a crisp label (e.g., transition regions between a foggy sky and the ground or between sand and water at a beach). In these cases, a visual word is best represented with partial memberships across multiple topics. To address this, we present a partial membership latent Dirichlet allocation (PM-LDA) model and associated parameter estimation algorithms. Experimental results on two natural image datasets and one SONAR image dataset show that PM-LDA can produce both crisp and soft semantic image segmentations; a capability existing methods do not have.
We propose Sentence Level Recurrent Topic Model (SLRTM), a new topic model that assumes the generation of each word within a sentence to depend on both the topic of the sentence and the whole history of its preceding words in the sentence. Different from conventional topic models that largely ignore the sequential order of words or their topic coherence, SLRTM gives full characterization to them by using a Recurrent Neural Networks (RNN) based framework. Experimental results have shown that SLRTM outperforms several strong baselines on various tasks. Furthermore, SLRTM can automatically generate sentences given a topic (i.e., topics to sentences), which is a key technology for real world applications such as personalized short text conversation.
A comprehensive artificial intelligence system needs to not only perceive the environment with different `senses' (e.g., seeing and hearing) but also infer the world's conditional (or even causal) relations and corresponding uncertainty. The past decade has seen major advances in many perception tasks such as visual object recognition and speech recognition using deep learning models. For higher-level inference, however, probabilistic graphical models with their Bayesian nature are still more powerful and flexible. In recent years, Bayesian deep learning has emerged as a unified probabilistic framework to tightly integrate deep learning and Bayesian models. In this general framework, the perception of text or images using deep learning can boost the performance of higher-level inference and in turn, the feedback from the inference process is able to enhance the perception of text or images. This survey provides a comprehensive introduction to Bayesian deep learning and reviews its recent applications on recommender systems, topic models, control, etc. Besides, we also discuss the relationship and differences between Bayesian deep learning and other related topics such as Bayesian treatment of neural networks. For a constantly updating project page, please refer to https://github.com/js05212/BayesianDeepLearning-Survey.
Identifying and communicating relationships between causes and effects is important for understanding our world, but is affected by language structure, cognitive and emotional biases, and the properties of the communication medium. Despite the increasing importance of social media, much remains unknown about causal statements made online. To study real-world causal attribution, we extract a large-scale corpus of causal statements made on the Twitter social network platform as well as a comparable random control corpus. We compare causal and control statements using statistical language and sentiment analysis tools. We find that causal statements have a number of significant lexical and grammatical differences compared with controls and tend to be more negative in sentiment than controls. Causal statements made online tend to focus on news and current events, medicine and health, or interpersonal relationships, as shown by topic models. By quantifying the features and potential biases of causality communication, this study improves our understanding of the accuracy of information and opinions found online.
In order to create a corpus exploration method providing topics that are easier to interpret than standard LDA topic models, here we propose combining two techniques called Entity linking and Labeled LDA. Our method identifies in an ontology a series of descriptive labels for each document in a corpus. Then it generates a specific topic for each label. Having a direct relation between topics and labels makes interpretation easier; using an ontology as background knowledge limits label ambiguity. As our topics are described with a limited number of clear-cut labels, they promote interpretability, and this may help quantitative evaluation. We illustrate the potential of the approach by applying it in order to define the most relevant topics addressed by each party in the European Parliament's fifth mandate (1999-2004).
Distributed dense word vectors have been shown to be effective at capturing token-level semantic and syntactic regularities in language, while topic models can form interpretable representations over documents. In this work, we describe lda2vec, a model that learns dense word vectors jointly with Dirichlet-distributed latent document-level mixtures of topic vectors. In contrast to continuous dense document representations, this formulation produces sparse, interpretable document mixtures through a non-negative simplex constraint. Our method is simple to incorporate into existing automatic differentiation frameworks and allows for unsupervised document representations geared for use by scientists while simultaneously learning word vectors and the linear relationships between them.
We propose an algorithm for detecting patterns exhibited by anomalous clusters in high dimensional discrete data. Unlike most anomaly detection (AD) methods, which detect individual anomalies, our proposed method detects groups (clusters) of anomalies; i.e. sets of points which collectively exhibit abnormal patterns. In many applications this can lead to better understanding of the nature of the atypical behavior and to identifying the sources of the anomalies. Moreover, we consider the case where the atypical patterns exhibit on only a small (salient) subset of the very high dimensional feature space. Individual AD techniques and techniques that detect anomalies using all the features typically fail to detect such anomalies, but our method can detect such instances collectively, discover the shared anomalous patterns exhibited by them, and identify the subsets of salient features. In this paper, we focus on detecting anomalous topics in a batch of text documents, developing our algorithm based on topic models. Results of our experiments show that our method can accurately detect anomalous topics and salient features (words) under each such topic in a synthetic data set and two real-world text corpora and achieves better performance compared to both standard group AD and individual AD techniques. All required code to reproduce our experiments is available from https://github.com/hsoleimani/ATD
Machines, not humans, are the world's dominant knowledge accumulators but humans remain the dominant decision makers. Interpreting and disseminating the knowledge accumulated by machines requires expertise, time, and is prone to failure. The problem of how best to convey accumulated knowledge from computers to humans is a critical bottleneck in the broader application of machine learning. We propose an approach based on human teaching where the problem is formalized as selecting a small subset of the data that will, with high probability, lead the human user to the correct inference. This approach, though successful for modeling human learning in simple laboratory experiments, has failed to achieve broader relevance due to challenges in formulating general and scalable algorithms. We propose general-purpose teaching via pseudo-marginal sampling and demonstrate the algorithm by teaching topic models. Simulation results show our sampling-based approach: effectively approximates the probability where ground-truth is possible via enumeration, results in data that are markedly different from those expected by random sampling, and speeds learning especially for small amounts of data. Application to movie synopsis data illustrates differences between teaching and random sampling for teaching distributions and specific topics, and demonstrates gains in scalability and applicability to real-world problems.
Recently, there has been considerable progress on designing algorithms with provable guarantees -- typically using linear algebraic methods -- for parameter learning in latent variable models. But designing provable algorithms for inference has proven to be more challenging. Here we take a first step towards provable inference in topic models. We leverage a property of topic models that enables us to construct simple linear estimators for the unknown topic proportions that have small variance, and consequently can work with short documents. Our estimators also correspond to finding an estimate around which the posterior is well-concentrated. We show lower bounds that for shorter documents it can be information theoretically impossible to find the hidden topics. Finally, we give empirical results that demonstrate that our algorithm works on realistic topic models. It yields good solutions on synthetic data and runs in time comparable to a {\em single} iteration of Gibbs sampling.
Topic models have emerged as fundamental tools in unsupervised machine learning. Most modern topic modeling algorithms take a probabilistic view and derive inference algorithms based on Latent Dirichlet Allocation (LDA) or its variants. In contrast, we study topic modeling as a combinatorial optimization problem, and propose a new objective function derived from LDA by passing to the small-variance limit. We minimize the derived objective by using ideas from combinatorial optimization, which results in a new, fast, and high-quality topic modeling algorithm. In particular, we show that our results are competitive with popular LDA-based topic modeling approaches, and also discuss the (dis)similarities between our approach and its probabilistic counterparts.
In retrospective assessments, internet news reports have been shown to capture early reports of unknown infectious disease transmission prior to official laboratory confirmation. In general, media interest and reporting peaks and wanes during the course of an outbreak. In this study, we quantify the extent to which media interest during infectious disease outbreaks is indicative of trends of reported incidence. We introduce an approach that uses supervised temporal topic models to transform large corpora of news articles into temporal topic trends. The key advantages of this approach include, applicability to a wide range of diseases, and ability to capture disease dynamics - including seasonality, abrupt peaks and troughs. We evaluated the method using data from multiple infectious disease outbreaks reported in the United States of America (U.S.), China and India. We noted that temporal topic trends extracted from disease-related news reports successfully captured the dynamics of multiple outbreaks such as whooping cough in U.S. (2012), dengue outbreaks in India (2013) and China (2014). Our observations also suggest that efficient modeling of temporal topic trends using time-series regression techniques can estimate disease case counts with increased precision before official reports by health organizations.
Probabilistic topic models are generative models that describe the content of documents by discovering the latent topics underlying them. However, the structure of the textual input, and for instance the grouping of words in coherent text spans such as sentences, contains much information which is generally lost with these models. In this paper, we propose sentenceLDA, an extension of LDA whose goal is to overcome this limitation by incorporating the structure of the text in the generative and inference processes. We illustrate the advantages of sentenceLDA by comparing it with LDA using both intrinsic (perplexity) and extrinsic (text classification) evaluation tasks on different text collections.
Recent advances in neural variational inference have spawned a renaissance in deep latent variable models. In this paper we introduce a generic variational inference framework for generative and conditional models of text. While traditional variational methods derive an analytic approximation for the intractable distributions over latent variables, here we construct an inference network conditioned on the discrete text input to provide the variational distribution. We validate this framework on two very different text modelling applications, generative document modelling and supervised question answering. Our neural variational document model combines a continuous stochastic document representation with a bag-of-words generative model and achieves the lowest reported perplexities on two standard test corpora. The neural answer selection model employs a stochastic representation layer within an attention mechanism to extract the semantics between a question and answer pair. On two question answering benchmarks this model exceeds all previous published benchmarks.
Music information retrieval faces a challenge in modeling contextualized musical concepts formulated by a set of co-occurring tags. In this paper, we investigate the suitability of our recently proposed approach based on a Siamese neural network in fighting off this challenge. By means of tag features and probabilistic topic models, the network captures contextualized semantics from tags via unsupervised learning. This leads to a distributed semantics space and a potential solution to the out of vocabulary problem which has yet to be sufficiently addressed. We explore the nature of the resultant music-based semantics and address computational needs. We conduct experiments on three public music tag collections -namely, CAL500, MagTag5K and Million Song Dataset- and compare our approach to a number of state-of-the-art semantics learning approaches. Comparative results suggest that this approach outperforms previous approaches in terms of semantic priming and music tag completion.
We present a token-level decision summarization framework that utilizes the latent topic structures of utterances to identify "summary-worthy" words. Concretely, a series of unsupervised topic models is explored and experimental results show that fine-grained topic models, which discover topics at the utterance-level rather than the document-level, can better identify the gist of the decision-making process. Moreover, our proposed token-level summarization approach, which is able to remove redundancies within utterances, outperforms existing utterance ranking based summarization methods. Finally, context information is also investigated to add additional relevant information to the summary.
This paper addresses the problem of summarizing decisions in spoken meetings: our goal is to produce a concise {\it decision abstract} for each meeting decision. We explore and compare token-level and dialogue act-level automatic summarization methods using both unsupervised and supervised learning frameworks. In the supervised summarization setting, and given true clusterings of decision-related utterances, we find that token-level summaries that employ discourse context can approach an upper bound for decision abstracts derived directly from dialogue acts. In the unsupervised summarization setting,we find that summaries based on unsupervised partitioning of decision-related utterances perform comparably to those based on partitions generated using supervised techniques (0.22 ROUGE-F1 using LDA-based topic models vs. 0.23 using SVMs).