Leveraging GPT-2 for Classifying Spam Reviews with Limited Labeled Data via Adversarial Training

TL;DR

This paper introduces an adversarial training approach using GPT-2 to improve spam review classification with limited labeled data, demonstrating superior accuracy and synthetic review generation capabilities.

Contribution

It presents a novel adversarial training framework leveraging GPT-2 for spam detection with scarce labeled data, enhancing performance and data augmentation.

Findings

Outperforms state-of-the-art methods by at least 7% accuracy with limited labels

Can generate realistic synthetic reviews to augment training data

Effective on TripAdvisor and YelpZip datasets

Abstract

Online reviews are a vital source of information when purchasing a service or a product. Opinion spammers manipulate these reviews, deliberately altering the overall perception of the service. Though there exists a corpus of online reviews, only a few have been labeled as spam or non-spam, making it difficult to train spam detection models. We propose an adversarial training mechanism leveraging the capabilities of Generative Pre-Training 2 (GPT-2) for classifying opinion spam with limited labeled data and a large set of unlabeled data. Experiments on TripAdvisor and YelpZip datasets show that the proposed model outperforms state-of-the-art techniques by at least 7% in terms of accuracy when labeled data is limited. The proposed model can also generate synthetic spam/non-spam reviews with reasonable perplexity, thereby, providing additional labeled data during training.