Classification of Fake News by Fine-tuning Deep Bidirectional Transformers based Language Model

Aggarwal, Akshay; Chauhan, Aniruddha; Kumar, Deepika; Mittal, Mamta; Verma, Sharad

doi:10.4108/eai.13-7-2018.163973

Cited by 22 publications

(16 citation statements)

References 28 publications

(28 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…BERT is composed of two stages, i.e., unsupervised pre-training and supervised fine-tuning. Aggarwal et al [5] showed that BERT outperformed LSTM and gradient boosted tree models even with minimal text pre-processing. To improve the performance of BERT, Jwa et al [27] proposed a model that classified the data using weighted cross-entropy.…”

Section: Related Workmentioning

confidence: 99%

“…Recently, more successful models based on deep learning techniques have been used to detect misinformation. For example, Aggarwal et al [5] detected misinformation using BERT with very minimal text pre-processing, but obtained very good performance. It was also reported that by April 2020 Facebook removed more than fifty million posts related to COVID-19 since they were classified as misinformation using machine learning-based NLP techniques [6].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Combat COVID-19 infodemic using explainable natural language processing models

Ayoub

Yang

Zhou

2021

Information Processing & Management

106

View full text Add to dashboard Cite

Misinformation of COVID-19 is prevalent on social media as the pandemic unfolds, and the associated risks are extremely high. Thus, it is critical to detect and combat such misinformation. Recently, deep learning models using natural language processing techniques, such as BERT (Bidirectional Encoder Representations from Transformers), have achieved great successes in detecting misinformation. In this paper, we proposed an explainable natural language processing model based on DistilBERT and SHAP (Shapley Additive exPlanations) to combat misinformation about COVID-19 due to their efficiency and effectiveness. First, we collected a dataset of 984 claims about COVID-19 with fact checking. By augmenting the data using back-translation, we doubled the sample size of the dataset and the DistilBERT model was able to obtain good performance

show abstract

Section: Related Workmentioning

confidence: 99%

mentioning

confidence: 99%

Combat COVID-19 infodemic using explainable natural language processing models

Ayoub

Yang

Zhou

2021

Information Processing & Management

106

View full text Add to dashboard Cite

show abstract

“…BERT is one of the newest advances in natural language modeling and is state-of-the-art in various text data sets ( Sun et al, 2019 ; Aggarwal et al, 2020 ; González-Carvajal and Garrido-Merchán, 2020 ).…”

Section: Methodsmentioning

confidence: 99%

Automatic classification of literature in systematic reviews on food safety using machine learning

Bulk

Bouzembrak

Gavai

et al. 2022

Current Research in Food Science

View full text Add to dashboard Cite

Systematic reviews are used to collect relevant literature to answer a research question in a way that is clear, thorough, unbiased and reproducible. They are implemented as a standard method in the domain of food safety to obtain a literature overview on the state-of-the-art research related to food safety topics of interest. A disadvantage to systematic reviews, however, is that this process is time-consuming and requires expert domain knowledge. The work reported here aims to reduce the time needed by an expert to screen all possible relevant articles by applying machine learning techniques to classify the articles automatically as either relevant or not relevant. Eight different machine learning algorithms and ensembles of all combinations of these algorithms were tested on two different systematic reviews on food safety (i.e. chemical hazards in cereals and leafy greens). The results showed that the best performance was obtained by an ensemble of naive Bayes and a support vector machine, resulting in an average decrease of 32.8% in the amount of articles the expert has to read and an average decrease in irrelevant articles of 57.8% while keeping 95% of the relevant articles. It was concluded that automatic classification of the literature in a systematic literature review can support experts in their task and save valuable time without compromising the quality of the review.

show abstract

“…The result, people are polarized into those who speaks positively or negatively [5]. There is a long way to go tackle the problem of fake news detection, transfer learning promises to be a strong means of progress in the field [6]. Furthermore, this article examines language abuse or language criminalization that appeared in the public sphere during the 2019 Indonesian Presidential election.…”

Section: Introduction Ministry Of Communication and Informationmentioning

confidence: 99%

Internet (Social Media) and Its Relationship with Hoax and Fake News in the 2019 Presidential Election

Syaputri¹,

Septianasari²,

Retnawati³

et al. 2021

Proceedings of the 1st Paris Van Java International Seminar on Health, Economics, Social Science and Humanities (PVJ-ISHESSH 20

View full text Add to dashboard Cite

This research observe the Indonesia presidential election 2019 focused on massive hoax and fake news on social media. These phenomena are giving such effects for society because it leads emergence chaos. The objective of this research identified fake news and described language misuse on Indonesia presidential election 2019 to reduce emergence chaos in society. Readers who have good competence will be able to distinguish between factual reports and opinion pieces that tend to be subjective. On the other hand, incompetent readers will be easily provoked by fictitious issues that tend to be provocative. Method of the research used descriptive qualitative on referential and phenomenology approaches. Data collected using scrutinized (simak) method. The data got from www.turnbackhoax.id website. This website is used as a data source because the data contained on this page is accompanied by valid evidence related to fake news or hoax that are discussed on social media. The data analysed by using theories from Yule (2006) and Huang (2006) about presupposition on landscape discourse. The result shown, misused refer to criminal action did by certain people in spreading provocative issues. Those issues propagate narration with picture to lead the public opinion.

show abstract

Classification of Fake News by Fine-tuning Deep Bidirectional Transformers based Language Model

Cited by 22 publications

References 28 publications

Combat COVID-19 infodemic using explainable natural language processing models

Combat COVID-19 infodemic using explainable natural language processing models

Automatic classification of literature in systematic reviews on food safety using machine learning

Internet (Social Media) and Its Relationship with Hoax and Fake News in the 2019 Presidential Election

Contact Info

Product

Resources

About