Generating Sentiment-Preserving Fake Online Reviews Using Neural Language Models and Their Human- and Machine-Based Detection

Adelani, David Ifeoluwa; Mai, Haotian; Fang, Fuming; Nguyen, Huy H.; Yamagishi, Junichi; Echizen, Isao

doi:10.1007/978-3-030-44041-1_114

Cited by 59 publications

(32 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The only research on the detection of deepfake social media texts was conducted by [5] on Amazon reviews written by GPT-2. They evaluated several human-machine discriminators: the Grover-based detector, GLTR, RoBERTa-based detector from OpenAI and a simple ensemble that fused these detectors using logistic regression at the score level.…”

Section: Plos Onementioning

confidence: 99%

TweepFake: About detecting deepfake tweets

et al. 2021

View full text Add to dashboard Cite

The recent advances in language modeling significantly improved the generative capabilities of deep neural models: in 2019 OpenAI released GPT-2, a pre-trained language model that can autonomously generate coherent, non-trivial and human-like text samples. Since then, ever more powerful text generative models have been developed. Adversaries can exploit these tremendous generative capabilities to enhance social bots that will have the ability to write plausible deepfake messages, hoping to contaminate public debate. To prevent this, it is crucial to develop deepfake social media messages detection systems. However, to the best of our knowledge no one has ever addressed the detection of machine-generated texts on social networks like Twitter or Facebook. With the aim of helping the research in this detection field, we collected the first dataset of real deepfake tweets, TweepFake. It is real in the sense that each deepfake tweet was actually posted on Twitter. We collected tweets from a total of 23 bots, imitating 17 human accounts. The bots are based on various generation techniques, i.e., Markov Chains, RNN, RNN+Markov, LSTM, GPT-2. We also randomly selected tweets from the humans imitated by the bots to have an overall balanced dataset of 25,572 tweets (half human and half bots generated). The dataset is publicly available on Kaggle. Lastly, we evaluated 13 deepfake text detection methods (based on various state-of-the-art approaches) to both demonstrate the challenges that Tweepfake poses and create a solid baseline of detection techniques. We hope that TweepFake can offer the opportunity to tackle the deepfake detection on social media messages as well.

show abstract

Section: Plos Onementioning

confidence: 99%

TweepFake: About detecting deepfake tweets

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Figure 4 shows an example of actual handwritten characters and the clones generated with the proposed method. [61] In addition to the modalities mentioned above, fluent sentences and paragraphs can also be automatically generated using recently developed deep-learning-based neural language models. Applications of neural language models mainly include machine translation, image captioning, text summarization, dialogue generation, and speech recognition.…”

Section: Handwritten Character Image Generation [60]mentioning

confidence: 99%

“…Here, we proposed a method for generating fake reviews that requires minimal skills but has high performance [61]. A user of this method can create a flood of positive or negative fake reviews to affect the rating of a product.…”

Section: Fake Review Generationmentioning

confidence: 99%

“…Testing showed that the proposed method could discriminate CGT generated using an RNN-based method [74] from HWT with an accuracy of 96.3%. [61] In Sect. 3.9, we introduced an easy-to-use but highperformance fake-review generation method.…”

Section: Passive Video Forgery Detection [71]mentioning

confidence: 99%

“…In addition, we can fuse these detectors by using logistic regression at the score level to further improve detection accuracy. Table 4 shows the detection results [61]. We used the method described in Sect.…”

Section: Fake Review Detectionmentioning

confidence: 99%

See 2 more Smart Citations