Counteracting French Fake News on Climate Change Using Language Models

Round 1

Reviewer 1 Report

The authors' work is original and timely. However, I have a few suggestions to improve the quality of this paper:

1. The result section of this paper is shallow. There should be enough visualizations (such as pie charts, bar charts, histograms, word clouds, etc.) to help readers interpret the results and appreciate the work done by the authors.

2. The author did not compare their results with existing work. It is hard to determine the novelty of the work. More so, it would have been much better if the authors can show novelty in the model used. It is not enough to use a different dataset and claim novelty.

3. The literature review of this work is not very comprehensive. For instance, this paper omits a relevant paper [Hong SC. Presumed effects of “fake news” on the global warming discussion in a cross-cultural context. Sustainability. 2020 Mar 9;12(5):2123.]

Author Response

Thank you kindly for your review and feedback, we highly appreciate it.

1. Thank you kindly for your suggestion. We have introduced confusion matrices and comparative distributions between true and predicted labels for the best models corresponding to the two tasks (i.e., article and sentence classifications). In addition, we have further detailed our results.
2. The main goal of the paper is to introduce our dataset and establish a strong baseline that helps assess the difficulty of the task and the quality of the dataset. Additional details have been added in the discussion section.
3. New references were discussed in the Introduction, including the one suggested by the reviewer.

Reviewer 2 Report

Authors presented a research work Counteracting French Fake News on Climate Change using Language Models. However, the contributions are limited at the current stage. The following changes can be incorporated to strengthen the quality of the manuscript:

1.      The structure and layout of the paper should be consistent throughout.

2.      The usage of the English Language should be improved throughout the paper. There are typos and grammatical errors in the text. I suggest rewriting some paragraphs. Native proofreading can improve the clarity of the manuscript.

3.      The abstract is not well-written. Be sure to write the abstract more briefly. Also, I suggest adding some information about the evaluation results.

4.      The Introduction section needs improvement. The authors need to clearly highlight the main aim and motivation of their analysis.

5.      Technical depth of the paper is limited and should be improved.

6.      Result section can be better presented along with a comparative analysis in the manuscript.

7.      Reference section should be strengthened with recent ones. Following references shouled be referred in the manuscript.

[1] - Fake News Classification Based on Content Level Features. Appl. Sci. 2022, 12, 1116. https://doi.org/10.3390/app12031116.

[2] - (2020). Digital Transparency and Open Data. In: Farazmand, A. (eds) Global Encyclopedia of Public Administration, Public Policy, and Governance. Springer, Cham. https://doi.org/10.1007/978-3-319-31816-5_3957-1

Author Response

Thank you kindly for your review and feedback, we highly appreciate it.

1. We checked and addressed the issues that we found. Please let us know if you spot anything else out of order.
2. We have performed a thorough check of the entire paper with a linguist.
3. The abstract has been rewritten in a more condensed manner.
4. Both the main aim and the motivation of our analysis were clearly specified in the Introduction of the article.
5. We have performed a thorough revision of the entire paper.
6. Multiple model architectures are compared in the results section, while confusion matrices and distribution plots have also been included.
7. The reference section was strengthened with more recent ones, including those suggested by the reviewer.

Round 2

Reviewer 2 Report

The suggested comments have been incorporated in the manuscript.