GraphTMT: Unsupervised Graph-based Topic Modeling from Video Transcripts

TL;DR

GraphTMT introduces an unsupervised, graph-based topic modeling method for video transcripts that enhances semantic coherence and does not require predefining the number of topics, outperforming baseline methods on real datasets.

Contribution

It presents a novel unsupervised graph clustering approach for topic extraction from video transcripts, eliminating the need to specify the number of topics beforehand.

Findings

Extracts coherent, meaningful topics from video transcripts

Outperforms baseline methods on MuSe-CaR dataset

Proves applicability on Citysearch corpus

Abstract

To unfold the tremendous amount of multimedia data uploaded daily to social media platforms, effective topic modeling techniques are needed. Existing work tends to apply topic models on written text datasets. In this paper, we propose a topic extractor on video transcripts. Exploiting neural word embeddings through graph-based clustering, we aim to improve usability and semantic coherence. Unlike most topic models, this approach works without knowing the true number of topics, which is important when no such assumption can or should be made. Experimental results on the real-life multimodal dataset MuSe-CaR demonstrates that our approach GraphTMT extracts coherent and meaningful topics and outperforms baseline methods. Furthermore, we successfully demonstrate the applicability of our approach on the popular Citysearch corpus.