Improving unsupervised neural aspect extraction for online discussions using out-of-domain classification

TL;DR

This paper proposes a sentence filtering method using out-of-domain classification to enhance the coherence of unsupervised neural aspect extraction in online discussions, especially for traditional text sources like news articles.

Contribution

It introduces a simple sentence filtering technique based on out-of-domain classification to improve neural aspect extraction without altering the core model architecture.

Findings

Sentence filtering improves topic coherence in newsgroup data

Filtering out low-probability in-domain sentences enhances aspect extraction quality

The method is effective across different traditional text sources

Abstract

Deep learning architectures based on self-attention have recently achieved and surpassed state of the art results in the task of unsupervised aspect extraction and topic modeling. While models such as neural attention-based aspect extraction (ABAE) have been successfully applied to user-generated texts, they are less coherent when applied to traditional data sources such as news articles and newsgroup documents. In this work, we introduce a simple approach based on sentence filtering in order to improve topical aspects learned from newsgroups-based content without modifying the basic mechanism of ABAE. We train a probabilistic classifier to distinguish between out-of-domain texts (outer dataset) and in-domain texts (target dataset). Then, during data preparation we filter out sentences that have a low probability of being in-domain and train the neural model on the remaining sentences. The positive effect of sentence filtering on topic coherence is demonstrated in comparison to aspect extraction models trained on unfiltered texts.