Abstract:In recent years, multimodal topic models have gained significant attention in various tasks involving short texts. Despite their impressive results, most models rely on bag-of-words assumptions for each modality, neglecting the intrinsic word relations in the textual modality and the underlying object relations in the visual modality. To address this limitation, we propose a novel approach that represents each document modality as a graph, harnessing the word relations and the visual object relations to guide the topic extraction process. Our approach is grounded in the insight that, in the textual modality, words with specific relations, such as co-occurrence relations, semantic relations and syntactic relations, are more likely to be assigned to the same topic. Similarly, in the visual modality, the relations between objects, such as spatial relations and contextual relations, can also provide valuable information for topic extraction. By leveraging graph-based representations, our model captures the inherent associations between words and visual objects, resulting in the generation of more coherent and interpretable topics. To infer the model's parameters, we develop an effective algorithm that integrates neural variational inference and contrastive learning. The experimental results on three datasets verify the effectiveness of our proposed model in terms of topic coherence, topic diversity and mean average precision, confirming that incorporating word relations and object relations through graph-based representations significantly enhances the quality of the extracted topics.

Identifying Objective and Subjective Words via Topic Modeling

Mining Coherent Topics in Documents Using Word Embeddings and Large-Scale Text Data

SBTM: A Joint Sentiment and Behaviour Topic Model for Online Course Discussion Forums

Bag-of-Discriminative-Words (BoDW) Representation via Topic Modeling.

Topic Models Incorporating Statistical Word Senses

Statistical Word Sense Aware Topic Models

Tackling topic general words in topic modeling.

Graph-based Multimodal Topic Modeling with Word Relations and Object Relations

Subjective Relation Extraction in Chinese Opinion Mining

Topic Modeling in Semantic Space with Keywords.

Jointly Discovering Fine-grained and Coarse-grained Sentiments Via Topic Modeling.

A Query-Driven Topic Model

Using Topic Modeling Methods for Short-Text Data: A Comparative Analysis

A Novel Topic Model for Documents by Incorporating Semantic Relations Between Words

Study language models with specific user goals.

Sao-Based Topic Modeling For Competitive Technical Intelligence: A Case Study In Graphene

BTM: Topic Modeling over Short Texts

Intensity of Relationship Between Words: Using Word Triangles in Topic Discovery for Short Texts

Using Word Sense As a Latent Variable in LDA Can Improve Topic Modeling.

Research on the Text Similarity Algorithms in Automatic Scoring of Subjective Questions