Author-topic Model Research Articles

An FMRI Meta-Analysis of Interoception in Eating Disorders

Eating Disorders (EDs) are associated with disturbed interoception – the sense of the internal condition of the body. Disturbances in interoception across senses have not yet been comprehensively examined in EDs. To do so, we employed an innovative Bayesian author-topic model approach to fMRI meta-analyses that pools together neural deficits across interoceptive senses and task types in participants with and recovered from EDs. Following PRISMA guidelines, our results combine activation patterns from 1,341 initially screened studies and data from 25 manuscripts that met study criteria that compare 463 patients with EDs (current or recovered) to 450 healthy control participants (HC). Altered brain activity was found within vision/sensory processing (precuneus), taste/self-referential processing (claustrum/posterior insula) and reward/set-shifting (global pallidus, medial frontal gyrus, anterior cingulate, precentral gyrus and parietal lobe) components in EDs compared to HC. Our results reveal separate components for bottom-up exteroceptive and interoceptive processing centering around the precuneus and claustrum/insula and also reward processing/set-shifting deficits. Thus, bottom-up sensory and reward processing are key deficits in EDs during ill and recovered states.

A Functional Magnetic Resonance Imaging Meta-Analysis of Childhood Trauma

BackgroundTraumatic experiences during childhood significantly impact the developing brain and contribute to the development of numerous physical and mental health problems. To date, however, a comprehensive understanding of the functional impairments within the brain associated with childhood trauma histories does not exist. Previous functional magnetic resonance imaging (fMRI) meta-analytical tools required homogeneity of task types and the clinical populations studied, thus preventing the comprehensive pooling of brain-based deficits present in children who have trauma histories. We hypothesized that the use of the novel, data-driven Bayesian author-topic model approach to fMRI meta-analyses would reveal deficits in brain networks that span fMRI task types in children with trauma histories. MethodsTo our knowledge, this is the first study to use the Bayesian author-topic model approach to fMRI meta-analyses within a clinical population. Using PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, we present data-driven results obtained by combining activation patterns across heterogeneous tasks from 1428 initially screened studies and combining data from 14 studies that met study criteria (285 children with trauma histories, 297 healthy control children). ResultsAltered brain activity was revealed within 2 clusters in children with trauma histories compared to control children: the default mode/affective network/posterior insula and the central executive network. Our identified clusters were associated with tasks pertaining to cognitive processing, emotional/social stress, self-referential thought, memory, unexpected stimuli, and avoidance behaviors in youths who have experienced childhood trauma. ConclusionsOur results reveal disturbances in children with trauma histories within the modulation of the default mode and central executive networks—but not the salience network—regardless of whether children also presented with posttraumatic stress symptoms.

Neural Embedding Allocation: Distributed Representations of Topic Models

Abstract We propose a method that uses neural embeddings to improve the performance of any given LDA-style topic model. Our method, called neural embedding allocation (NEA), deconstructs topic models (LDA or otherwise) into interpretable vector-space embeddings of words, topics, documents, authors, and so on, by learning neural embeddings to mimic the topic model. We demonstrate that NEA improves coherence scores of the original topic model by smoothing out the noisy topics when the number of topics is large. Furthermore, we show NEA’s effectiveness and generality in deconstructing and smoothing LDA, author-topic models, and the recent mixed membership skip-gram topic model and achieve better performance with the embeddings compared to several state-of-the-art models.

An improved author-topic (AT) model with authorship credit allocation schemes

Authorship credit allocation schemes have attracted considerable research attention. However, no consensus about which one is the best has been attained until now, and limited evidence from practical tasks has been reported. Therefore, this study uses the author interest discovery task as a real-world task case to provide valuable insights into authorship credit allocation schemes and guidelines for further practical applications. For this purpose, a novel model, ATcredit, is proposed to strengthen the Author-Topic (AT) model with an authorship credit allocation scheme, and collapsed Gibbs sampling is used to approximate the posterior and estimate model parameters. Extensive experiments using the SynBio dataset reveal several interesting findings as follows. (a) Any scheme for allocating unequal authorship credits performs better than its equal-credit counterpart with our ATcredit model in terms of perplexity. (b) The fixed versions of four out of the six schemes work better than their flexible counterparts with our ATcredit model, regardless of the hyper-authorship strategy. (c) The variation coefficient of credit awards can serve as a criterion to decide whether the hyper-authorship strategy should be used. (d) When the number of authors in a scholarly article is less than three, the six authorship credit allocation schemes are similar to each other with our ATcredit model in terms of perplexity. (e) The harmonic counting scheme performs the best, followed by the arithmetic counting scheme, and the network-based counting scheme performs the worst with our ATcredit model in terms of perplexity. (f) The arithmetic counting scheme is similar to the harmonic counting scheme in terms of the normalised mutual information (NMI) of discovered interests, but the geometric counting scheme is different from the axiomatic and network-based counting schemes.

Collaboration of issuing agencies and topic evolution of health informatisation policies in China

Digital transformation in the Chinese healthcare industry has led national government agencies to issue a series of policies to guide the construction of health informatisation. However, little is known about the issuing agencies and the topics of health informatisation policies. This study aimed to explore the collaboration of policies issuing and the evolution of policy topics. In this study, a total of 156 policy documents were identified. Author–Topic model and pre-discretised method based on Latent Dirichlet Allocation model were employed to mine the correlation between the issuing agencies and policy topics, and the evolution of policy contents. Findings suggest that the development of health informatisation policies can be divided into three stages. The number of policies has been increasing constantly, among which the policy of opinion and notification accounts for the vast majority. Many government agencies are involved in formulating policies collaboratively. On the whole, the topics changed constantly over time. From 2003 to 2008, policy topics focused on standards and specifications, with the phenomenon of splitting and development. From 2009 to 2014, policies were predominantly related to the construction of regional health informatisation, with some emerging topics generating. Internet + medical and new information technology gained attention from 2015 to 2020; most topics in this period were inherited, split or merged from the previous period. This study is helpful to research and formulation of the health informatisation-related policies.

Topic Modeling in Twitter: Aggregating Tweets by Conversations

We propose a new pooling technique for topic modeling in Twitter, which groups together tweets occurring in the same user-to-user conversation. Under this scheme, tweets and their replies are aggregated into a single document and the users who posted them are considered co-authors. To compare this new scheme against existing ones, we train topic models using Latent Dirichlet Allocation (LDA) and the Author-Topic Model (ATM) on datasets consisting of tweets pooled according to the different methods. Using the underlying categories of the tweets in this dataset as a noisy ground truth, we show that this new technique outperforms other pooling methods in terms of clustering quality and document retrieval.

Author topic model for co-occurring normal documents and short texts to explore individual user preferences

The investigation of user preferences through user comments has attracted significant attention. Although topic models have been verified as useful tools to facilitate the understanding of textual contents, they cannot be directly applied to accomplish this task because of two problems. The first problem is the severe data sparsity suffered by user comments because they are generally short. The second problem is the mixture of opinions from both user comments and the original documents the users commented on. To simultaneously solve the data sparsity problem and explore clean user preferences, we propose an author co-occurring topic model (AOTM) for normal documents and their short user comments. By considering authorship, AOTM allows each author of short texts to have a probability distribution over a set of topics represented only short texts. Accordingly, the individual user preferences can be investigated based on these author-level distributions. We verify the performance of AOTM using two news article datasets and one e-commerce dataset. Extensive experiments demonstrate that the AOTM outperforms several state-of-the-art methods in topic learning and topic representation of documents. The potential usage of AOTM in exploring individual user preferences is further illustrated by drawing user portraits and predicting user posting behaviors.

Joint learning of author and citation contexts for computing drift in scholarly documents

Scholarly documents are sources of information on research topics written by academic experts. Topic drift in such scholarly documents is usually linked with the contextual variation in the title or abstract or entire document over time. However, topic distribution over words in different components of the document is non-uniform due to the varying impact of authors and citations, and their contribution to drift must be processed accordingly. This paper builds a model that distinguishes the context of a research document based on the author and citation by incorporating relation between topic, author, citation, word and time in the form of author context vector and citation context vector. To infer posterior probabilities, a parallel author cited_author topic model is presented. Continuous time bivariate Brownian motion model is employed for deducing the evolving bivariate topic parameters, specific to the author and citation. The word, topic pairs from the author and citation context vectors are jointly learned to yield topical word embeddings over time conditioned on author and citation contexts. When evaluated with NIPS and business journals datasets, the proposed model identifies topical variations over time precisely compared to other methods. It is found that broadening of topic happens due to the author context, and topic deviation is mainly caused by citation context.

Clustering student Instagram accounts using author-topic model

This study proposes topic model to cluster a group of high school teenager's Instagram account in Surabaya, Indonesia by using the author-topic models method. We collect valid 235 Instagram accounts (133 female, 102 male students). We gather a total of 3,346 captions of Instagram posts from 18 senior high schools. We find topics that define their Instagram's post or caption; these seven topics are namely: feeling, Surabaya events, photography, artists, vacation, religion and music. Through the process, the lowest perplexity come from 90 iterations, which suggests six groups of topics. The six topics are concluded based on the lowest perplexity value and labelled according to the words included in the topic. The topic of photography discussed by six schools. Photography, artists and vacation are discussed by three schools, while feeling and religion and music are being discussed by two and one school respectively.

Venue Topic Model–enhanced Joint Graph Modelling for Citation Recommendation in Scholarly Big Data

Natural language processing technologies, such as topic models, have been proven to be effective for scholarly recommendation tasks with the ability to deal with content information. Recently, venue recommendation is becoming an increasingly important research task due to the unprecedented number of publication venues. However, traditional methods focus on either the author’s local network or author-venue similarity, where the multiple relationships between scholars and venues are overlooked, especially the venue–venue interaction. To solve this problem, we propose an author topic model–enhanced joint graph modeling approach that consists of venue topic modeling, venue-specific topic influence modeling, and scholar preference modeling. We first model the venue topic with Latent Dirichlet Allocation. Then, we model the venue-specific topic influence in an asymmetric and low-dimensional way by considering the topic similarity between venues, the top-influence of venues, and the top-susceptibility of venues. The top-influence characterizes venues’ capacity of exerting topic influence on other venues. The top-susceptibility captures venues’ propensity of being topically influenced by other venues. Extensive experiments on two real-world datasets show that our proposed joint graph modeling approach outperforms the state-of-the-art methods.

Uncovering research trends and topics of communities in machine learning

This paper aims to uncover the research topics in machine learning research communities in a scientific collaboration network (SCN) to enhance the characteristic of systems such as retrieval or recommendation in intelligence-based systems. The existing research mainly focuses on the community evolution and measurement of typical features of the network. It is however unexplored how to identify the research interest of the communities along with authors in each community. A dataset is prepared consisting of 21,906 scientific articles from six top journals in the field of machine learning published from 1988 to 2017. An integrated approach combining the author-topic (AT) model with communities using through the directed affiliations (CoDA) method is explored to identify the research interest of the communities in a scientific collaboration network. The top rank communities are identified using the crank network community prioritization method. Finally, the similarity and dissimilarity of research interest in communities across decades are uncovered using the cosine similarity. The experimental results demonstrate the effectiveness and efficacy of the proposed technique. This study may be helpful for upcoming researchers to explore the research trends and topics in machine learning research communities.

Research on Scholar Clustering and Academic Influence Evaluation Method Based on Author Topic Model

Research on Scholar Clustering and Academic Influence Evaluation Method Based on Author Topic Model

Beyond consensus: Embracing heterogeneity in curated neuroimaging meta-analysis

Coordinate-based meta-analysis can provide important insights into mind-brain relationships. A popular approach for curated small-scale meta-analysis is activation likelihood estimation (ALE), which identifies brain regions consistently activated across a selected set of experiments, such as within a functional domain or mental disorder. ALE can also be utilized in meta-analytic co-activation modeling (MACM) to identify brain regions consistently co-activated with a seed region. Therefore, ALE aims to find consensus across experiments, treating heterogeneity across experiments as noise. However, heterogeneity within an ALE analysis of a functional domain might indicate the presence of functional sub-domains. Similarly, heterogeneity within a MACM analysis might indicate the involvement of a seed region in multiple co-activation patterns that are dependent on task contexts. Here, we demonstrate the use of the author-topic model to automatically determine if heterogeneities within ALE-type meta-analyses can be robustly explained by a small number of latent patterns. In the first application, the author-topic modeling of experiments involving self-generated thought (N = 179) revealed cognitive components fractionating the default network. In the second application, the author-topic model revealed that the left inferior frontal junction (IFJ) participated in multiple task-dependent co-activation patterns (N = 323). Furthermore, the author-topic model estimates compared favorably with spatial independent component analysis in both simulation and real data. Overall, the results suggest that the author-topic model is a flexible tool for exploring heterogeneity in ALE-type meta-analyses that might arise from functional sub-domains, mental disorder subtypes or task-dependent co-activation patterns. Code for this study is publicly available (https://github.com/ThomasYeoLab/CBIG/tree/master/stable_projects/meta-analysis/Ngo2019_AuthorTopic).

Clustering Student Instagram accounts using Author-Topic Model Based

This study proposes topic model to cluster a group of high school teenager's Instagram account in Surabaya, Indonesia by using the author-topic models method. We collect valid 235 Instagram accounts (133 female, 102 male students). We gather a total of 3,346 captions of Instagram posts from 18 senior high schools. We find topics that define their Instagram's post or caption; these seven topics are namely: feeling, Surabaya events, photography, artists, vacation, religion and music. Through the process, the lowest perplexity come from 90 iterations, which suggests six groups of topics. The six topics are concluded based on the lowest perplexity value and labelled according to the words included in the topic. The topic of photography discussed by six schools. Photography, artists and vacation are discussed by three schools, while feeling and religion and music are being discussed by two and one school respectively.

Discovering author interest evolution in order-sensitive and Semantic-aware topic modeling

Modeling the interests of authors over time from documents has important applications in broad applications such as recommendation systems, authorship identification and opinion extraction. In this paper, we propose an Ordering-sensitive and Semantic-aware Dynamic Author Topic Model (OSDATM), which monitors the evolution of author interest in time-stamped documents. The model further uses the discovered author interest information to discover better topics. Unlike traditional topic models, OSDATM is sensitive to the ordering of words, thus it extracts more information from the semantic meaning of the context. The experimental results show that OSDATM learns better topics than state-of-the-art topic models. In addition, the dynamic interests of authors that the OSDATM model discovers are interpretable and consistent with the truth.

Information Organization Patterns from Online Users in a Social Network

Recent years have seen the rise of user-generated contents (UGCs) in online social media. Diverse UGC sources and information overload are making it increasingly difficult to satisfy personalized information needs. To organize UGCs in a user-centered way, we should not only map them based on textual topics but also link them with users and even user communities. We propose a multi-dimensional framework to organize information by connecting UGCs, users, and user communities. First, we use a topic model to generate a topic hierarchy from UGCs. Second, an author-topic model is applied to learn user interests. Third, user communities are detected through a label propagation algorithm. Finally, a multi-dimensional information organization pattern is formulated based on similarities among the topic hierarchies of UGCs, user interests, and user communities. The results reveal that: 1) our proposed framework can organize information from multiple sources in a user-centered way; 2) hierarchical topic structures can provide comprehensive and in-depth topics for users; and, 3) user communities are efficient in helping people to connect with others who have similar interests.

Sustainable Computing Based Deep Learning Framework for Writing Research Manuscripts

Writing research manuscripts is always a tough task at the eleventh hour. Often researchers do not find time to rewrite the manuscript to satisfaction, which is not quantifiable though. This paper proposes a sustainable computing based deep learning framework for iterated accumulation of ideas while writing research manuscripts. The framework suggests Deep Author Topic Models (DATM) where every author of the manuscript is modeled. For this, we have assumed time based sustainable computing as a measure of evaluation for research manuscript effectiveness. Using respective DATM, the region contributed by every author in the manuscriptis analyzed and fine-tuned semantically such that the manuscript is made to perfection in least time.

Author–Subject–Topic model for reviewer recommendation

Interdisciplinary studies are becoming increasingly popular, and research domains of many experts are becoming diverse. This phenomenon brings difficulty in recommending experts to review interdisciplinary submissions. In this study, an Author–Subject–Topic (AST) model is proposed with two versions. In the model, reviewers’ subject information is embedded to analyse topic distributions of submissions and reviewers’ publications. The major difference between the AST and Author–Topic models lies in the introduction of a ‘Subject’ layer, which supervises the generation of hierarchical topics and allows sharing of subjects among authors. To evaluate the performance of the AST model, papers in Information System and Management (a typical interdisciplinary domain) in a famous Chinese academic library are investigated. Comparative experiments are conducted, which show the effectiveness of the AST model in topic distribution analysis and reviewer recommendation for interdisciplinary studies.

Latent Topic-based Subspace for Natural Language Processing

Natural Language Processing (NLP) applications have difficulties in dealing with automatically transcribed spoken documents recorded in noisy conditions, due to high Word Error Rates (WER), or in dealing with textual documents from the Internet, such as forums or micro-blogs, due to misspelled or truncated words, bad grammatical form… To improve the robustness against document errors, hitherto-proposed methods map these noisy documents in a latent space such as Latent Dirichlet Allocation (LDA), supervised LDA and author-topic (AT) models. In comparison to LDA, the AT model considers not only the document content (words), but also the class related to the document. In addition to these high-level representation models, an original compact representation, called c-vector, has recently been introduced avoid the tricky choice of the number of latent topics in these topic-based representations. The main drawback in the c-vector space building process is the number of sub-tasks required. Recently, we proposed both improving the performance of this c-vector compact representation of spoken documents and reducing the number of needed sub-tasks, using an original framework in a robust low dimensional space of features from a set of AT models called “Latent Topic-based Subspace” (LTS). This paper goes further by comparing the original LTS-based representation with the c-vector technique as well as with the state-of-the-art compression approach based on neural networks Encoder-Decoder (Autoencoder) and classification methods called deep neural networks (DNN) and long short-term memory (LSTM), on two classification tasks using noisy documents taking the form of speech conversations but also with textual documents from the 20-Newsgroups corpus. Results show that the original LTS representation outperforms the best previous compact representations with a substantial gain of more than 2.1 and 3.3 points in terms of correctly labeled documents compared to c-vector and Autoencoder neural networks respectively. An optimization algorithm of the scoring model parameters is then proposed to improve both the robustness and the performance of the proposed LTS-based approach. Finally, an automatic clustering approach based on the radial proximity between documents classes is introduced and shows promising performances.

Discovering cross-topic collaborations among researchers by exploiting weighted association rules

Identifying the most relevant scientific publications on a given topic is a well-known research problem. The Author-Topic Model (ATM) is a generative model that represents the relationships between research topics and publication authors. It allows us to identify the most influential authors on a particular topic. However, since most research works are co-authored by many researchers the information provided by ATM can be complemented by the study of the most fruitful collaborations among multiple authors. This paper addresses the discovery of research collaborations among multiple authors on single or multiple topics. Specifically, it exploits an exploratory data mining technique, i.e., weighted association rule mining, to analyze publication data and to discover correlations between ATM topics and combinations of authors. The mined rules characterize groups of researchers with fairly high scientific productivity by indicating (1) the research topics covered by their most cited publications and the relevance of their scientific production separately for each topic, (2) the nature of the collaboration (topic-specific or cross-topic), (3) the name of the external authors who have (occasionally) collaborated with the group either on a specific topic or on multiple topics, and (4) the underlying correlations between the addressed topics. The applicability of the proposed approach was validated on real data acquired from the Online Mendelian Inheritance in Man catalog of genetic disorders and from the PubMed digital library. The results confirm the effectiveness of the proposed strategy.