Generating Sentiment-Preserving Fake Online Reviews Using Neural Language Models and Their Human- and Machine-based Detection
arXiv:1907.09177
Abstract
Advanced neural language models (NLMs) are widely used in sequence generation tasks because they are able to produce fluent and meaningful sentences. They can also be used to generate fake reviews, which can then be used to attack online review systems and influence the buying decisions of online shoppers. To perform such attacks, it is necessary for experts to train a tailored LM for a specific topic. In this work, we show that a low-skilled threat model can be built just by combining publicly available LMs and show that the produced fake reviews can fool both humans and machines. In particular, we use the GPT-2 NLM to generate a large number of high-quality reviews based on a review with the desired sentiment and then using a BERT based text classifier (with accuracy of 96%) to filter out reviews with undesired sentiments. Because none of the words in the review are modified, fluent samples like the training data can be generated from the learned distribution. A subjective evaluation with 80 participants demonstrated that this simple method can produce reviews that are as fluent as those written by people. It also showed that the participants tended to distinguish fake reviews randomly. Three countermeasures, Grover, GLTR, and OpenAI GPT-2 detector, were found to be difficult to accurately detect fake review.
arXiv commentThe 34-th International Conference on Advanced Information Networking and Applications (AINA-2020)
References in corpus (11)
- Character-Aware Neural Language Models
- Deep Text Classification Can be Fooled
- Learning to Generate Reviews and Discovering Sentiment
- Release Strategies and the Social Impacts of Language Models
- Defending Against Neural Fake News
- Multiplicative LSTM for sequence modelling
- Neural Text Generation: Past, Present and Beyond
- Toward Diverse Text Generation with Inverse Reinforcement Learning
- Character-Word LSTM Language Models
- Large Scale Language Modeling: Converging on 40GB of Text in Four Hours
- The Dark Side of Micro-Task Marketplaces: Characterizing Fiverr and Automatically Detecting Crowdturfing