Pith. sign in

REVIEW 5 major objections 6 minor 28 references

Network-informed Prompt Engineering against Organized Astroturf Campaigns under Extreme Class Imbalance

T0 review · 5 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A frozen LLM, prompted with the tweet text, a text-encoded retweet tree, and balanced similar/opposite examples, can identify coordinated astroturf campaigns without any training or fine-tuning, outperforming graph neural network…

desk verdict A genuinely interesting frozen-LLM + balanced RAG idea for astroturf detection, but the evaluation likely suffers from train/test leakage and the headline gains are overstated. read the letter →

arxiv 2501.11849 v3 pith:O4DHV2M3 submitted 2025-01-21 cs.CL cs.AIcs.SI

Nikos Kanakaris , Heng Ping , Xiongye Xiao , Nesreen K. Ahmed , Luca Luceri , Emilio Ferrara , Paul Bogdan This is my paper · ORCID
classification cs.CLcs.AIcs.SI
keywords astroturfcampaignsdisinformationdetectionfakenewslargelanguagemodelspromptengineeringretrieval-augmentedgenerationclassimbalancegraphencoding
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims that a frozen large language model, fed with prompt-engineered text that combines the original tweet, a textual encoding of its retweet propagation tree, and a balanced set of retrieved examples, can detect coordinated political fake-news campaigns (astroturfing) on X better than graph neural network baselines. The central novelty is Balanced RAG: instead of retrieving only the most similar examples, the method retrieves similar examples and pairs each with the most similar example of the opposite label, so the LLM sees contrastive pairs that keep it attentive to the rare positive class. If this is right, astroturf detection can work without training or fine-tuning any model, relying only on off-the-shelf LLM inference and prompt design. On a dataset of over 10,000 propagation trees from the 2016 US election, the best variant reaches about 0.42 precision, 0.85 recall, 0.57 F1, and 0.86 ROC AUC, roughly a 2x-3x improvement over the GNN baselines tested.

What carries the argument

The load-bearing pieces are three: (1) propagation-tree reconstruction, which turns X's star-shaped retweet data into a tree using two heuristic conditions and, when neither applies, a random parent selected with probability proportional to the author's follower count (Eq. 1); (2) graph-to-text encoding, which renders the edge set as strings like '(2->1), (3->1)' so a text-only LLM can reason about structure; and (3) Balanced RAG, which retrieves the top-n training propagation trees most similar to the query by cosine similarity of embeddings, then for each retrieved tree finds the most similar tree with the opposite label, forming contrastive pairs that are placed into the few-shot prompt. The interaction of these components, not any single one, is what the paper claims carries the detection performance.

What would settle it

Run the best variant (Llama 3.1 70B + Balanced RAG) on the same test split but substitute the graph text with (a) the original X star graph (all retweets attached to the root) and (b) a degree-matched randomly rewired tree; if F1 stays at or above 0.55 in both cases, the reconstructed propagation-tree structure is not carrying the performance.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that graph structure does not need a learned encoder: a frozen decoder-only LLM can use a text encoding of a propagation tree, together with balanced few-shot examples, to identify coordinated disinformation campaigns under extreme class imbalance, outperforming GNN baselines by a factor of two to three in precision, recall, and F1. The authors assert that the combination of role prompting, chain-of-thought, and few-shot prompting, with Balanced RAG constructing pairs of similar-but-oppositely-labeled propagation trees, is what lets the LLM overcome the scarcity of positive samples without data augmentation or fine-tuning.

Load-bearing premise

The reconstructed propagation trees are assumed to faithfully reflect the true retweet cascades, even though X provides only star graphs; the paper builds trees with two heuristic rules and a follower-count-weighted random parent, so if those inferred edges are mostly wrong, the graph information in every prompt is largely noise.

Editorial extensions

If this is right

  • Astroturf detection becomes a zero-training task: any decoder-only LLM can be pointed at a tweet and its retweet cascade, and the balanced few-shot prompts supply the task knowledge.
  • The Balanced RAG pairing recipe transfers to other imbalanced few-shot classification problems, such as fraud detection or rare-disease mention classification, wherever similar-looking items carry opposite labels.
  • Text-encoding graphs decouples graph analysis from GNN tooling: propagation trees can be processed by any text model, making the approach portable across LLMs and easy to update as models change.
  • Because the best variant reports ~0.85 recall on a dataset where only 9.7% of trees are positive, the framework is viable as a first-pass filter that flags suspected campaigns for human review, trading precision for coverage.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A likely unstated reason for the large recall gain is that Balanced RAG acts as a label-balancing prior, nudging the LLM away from the majority-class bias; a testable variant would swap the retrieved contrastive pairs for randomly sampled balanced pairs matched on text length and centrality statistics to see how much of the gain is due to the similarity-based pairing specifically.
  • If the reconstructed trees are largely wrong, the graph text may function mainly as a mild regularizer, with most signal coming from tweet text and the retrieved examples; replacing the graph text with the raw star graph, or with a degree-matched random tree, would reveal whether the specific edge structure matters.
  • The framework inherits the biases of its training labels: the 2016-election dataset's 'positive' label comes from PolitiFact verdicts, so on other topics or platforms the same prompts would need new balanced example pools, and the paper does not test cross-domain transfer.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. This paper proposes a framework for detecting coordinated 'astroturf' political disinformation campaigns on X using a frozen large language model. The pipeline reconstructs a propagation tree from retweet data, encodes the tree structure as text, optionally retrieves labeled similar examples from a training set by embedding similarity, and queries Llama 3.1 with a prompt that combines the tweet text, graph encoding, and retrieved examples. The two main contributions are a graph-as-text prompting scheme and 'Balanced RAG,' which augments the prompt with both positive and negative retrieved examples to mitigate extreme class imbalance. On a FakeNewsNet-derived dataset of 10,228 propagation trees (9.7% positive), the best variant reports precision 0.4238, recall 0.8507, F1 0.5658, and ROC AUC 0.8602, outperforming GAT, GCN, GraphSAGE, and GraphConv baselines.

Significance. If the reported numbers survive a leakage-free evaluation, the framework would be a useful, training-free alternative to GNN-based fake-news detection, and the Balanced RAG idea—providing contrasting positive and negative demonstrations at inference time—is simple and potentially transferable to other class-imbalance settings. The paper ships source code, provides a full prompt template in the appendix, and uses a frozen LLM, which aids reproducibility and lowers computational barriers. However, the current evidence is not sufficient: the evaluation protocol leaves open a text-level train/test leakage route, the baselines are not class-imbalance-aware, and the headline '2x-3x' claim is not supported by the reported numbers. The core idea is worth pursuing, but the empirical claims need to be re-established under a claim-disjoint split and a fairer baseline setup.

major comments (5)
  1. [§4.1, §3.5, §4.3] The paper never describes the train/test split. The dataset in §4.1 is derived from FakeNewsNet, where labels are attached to underlying news articles or claims, and Balanced RAG in §3.5 retrieves the most text-similar propagation trees from the training set and places their texts and labels in the prompt. If the split is not claim-disjoint, a test tweet from a fake-news article can retrieve a near-identical positive training example and the LLM only needs to copy the label. Table 6 shows that removing text collapses recall from 0.8507 to 0.0622, so text is the decisive channel; this makes text-level leakage a decisive confound for every reported comparison. Please specify the split, ensure no claim or article overlap between train and test, and report results on a claim-disjoint split together with examples of retrieved neighbors.
  2. [Abstract and Table 2] The abstract claim of '2x-3x improvements in terms of precision, recall and F1 scores' is not supported by Table 2. The best recall is 0.8507 versus 0.5500 for GAT, a 1.55x improvement; F1 and precision improvements are about 2.45x and 2.33x, respectively. The claim that the framework achieves 'state-of-the-art performance' is also unsupported because no prior LLM-based or class-imbalance-aware methods are compared. Please revise the claims to match the reported numbers and restrict comparisons to the baselines actually tested.
  3. [§3.6, §4.5] Equation (6) formally defines the objective as maximizing the average score over the test set using the correct answers a_i, which is an oracle-based prompt selection procedure. The paper does not describe a validation protocol for choosing the number of RAG examples, the temperature, the retrieval embedding, or the graph encoding scheme, and the hyperparameter experiments in §4.5 use Llama 3.1 8B while the main results use 70B. No repeated runs, standard deviations, or significance tests are reported. Please move prompt and component selection to a validation set and report variance across runs or seeds.
  4. [§3.3, §4.4] The GNN baselines are trained with ordinary cross-entropy on a dataset with 9.7% positives and are not equipped with class weights, oversampling, or threshold tuning, whereas the proposed LLM variant receives balanced positive and negative demonstrations by construction. This makes the comparison lopsided; adding imbalance-aware baselines or balanced-loss and oversampled variants is needed. Separately, the propagation-tree reconstruction in §3.3 uses a random parent chosen with probability proportional to follower count (Eq. 1) when the two conditions fail; since X provides only a star graph, the reconstructed trees may be largely noise, and the paper does not analyze sensitivity to this randomness or validate the reconstruction. Either validation or a sensitivity analysis is required before attributing gains to graph information.
  5. [§3.5, §4.3] The definition of Balanced RAG in §3.5 (retrieve top-n similar samples, then for each retrieve the most similar opposite-label sample) differs from the implementation described in §4.3 (retrieve three most similar positives and three most similar negatives). These procedures are not equivalent and can produce different prompts; please align the description with the code and clarify which pairs are actually fed to the LLM, including whether the query appears in the retrieved set.
minor comments (6)
  1. [§4.3] The parenthetical labels are swapped: 'positive (organic propagation trees)' and 'three negative (astroturf propagation trees)' should read 'positive (astroturf propagation trees)' and 'three negative (organic propagation trees)'.
  2. [§4.1] There is a typo: 'propagation tressG' should be 'propagation trees G'.
  3. [Appendix B] The statement that the framework 'eliminates the need for labeled datasets' contradicts the use of a labeled training set for Balanced RAG and few-shot examples; please rephrase.
  4. [Tables 2, 5, 6] The captions of Tables 2, 5, and 6 duplicate the sentence 'The best results for each metric are highlighted in bold'; remove the duplicate.
  5. [Appendix B and §4.4] There are minor language errors: 'astoturf' should be 'astroturf' in Appendix B, and 'represent a 1.5x-3x improvement' should be 'represents a 1.5x-3x improvement' in §4.4.
  6. [§2] The related-work discussion of coordinated disinformation detection describes the present framework's contributions as established literature without citations; please distinguish prior work from this paper or add appropriate attributions.

Circularity Check

1 steps flagged · score 6.0 of 10

Reported gains are optimized on the test set via Eq. (6), making the headline improvement a fitted maximum rather than an out-of-sample prediction; no other step reduces to its inputs by construction.

  1. fitted input called prediction [Section 3.6, Eq. (6)]
    "We aim to find the best prompt combination such that the average performance of the model f over the test set S of the dataset is maximized with respect to a predefined metric: max_x 1/|S| sum_{i=1}^{|S|} score_f(g_i, W_{g_i}, t_{1_i}, a_i)"

    This equation makes the test set S the objective for selecting the prompt configuration. The paper then reports the best-performing variant (Llama 70B + Balanced RAG) from a menu of zero-shot, few-shot, RAG, and Balanced RAG settings, with additional sweeps over temperature, example count, and model size. Because the same test set is used both to choose the prompt and to compute the final precision, recall, and F1 scores, the reported 2x-3x improvement over baselines is an optimized maximum over test labels, not the performance of a fixed method on unseen data. The 'prediction' is therefore fitted to the test set by construction.

full rationale

The only clear circularity in the paper is in the evaluation protocol. Section 3.6 explicitly defines the goal as maximizing the average score over the test set S, and the experimental sections report the best configuration found by trying multiple prompt and RAG variants. This makes the headline numbers an optimized quantity rather than an out-of-sample evaluation, which fits the fitted-input-called-prediction pattern. The rest of the framework is not circular: the frozen LLM is not trained, Balanced RAG uses training labels only as few-shot exemplars, and the propagation-tree reconstruction is an imported modeling assumption (citing both self and external work) rather than a definitional equivalence. The potential text-level train/test leakage from retrieval is a serious correctness risk, but it is not a circularity because the paper does not describe the split and no equation forces the leakage. The self-citation to the tree-reconstruction method is load-bearing but not uniquely so, as an external citation is also present, so it does not by itself raise the score above the evaluation circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The framework has no trainable parameters, but it depends on several hand-chosen hyperparameters and strong domain assumptions. The retrieval embeddings, number of examples, temperature, and graph encoding are picked by the authors. The central assumptions are that the reconstructed propagation trees are accurate, that cosine similarity over the chosen embeddings captures coordination-relevant similarity, and that a frozen LLM can compare text-encoded cascades reliably. No invented physical or conceptual entities are introduced.

free parameters (4)
  • number of RAG examples = 6
    Selected via hyperparameter tuning (Table 4) on Llama 3.1 8B; 6 examples are used in the final 70B experiments.
  • temperature = 0
    Table 3 shows temperature 0 yields the best F1 and ROC AUC on the 8B model; chosen for determinism.
  • retrieval embedding representation = not specified (bag-of-words or pre-trained LLM)
    Section 3.5 leaves the embedding choice open, but the Balanced RAG results depend on this choice and no specific implementation is reported.
  • graph encoding scheme = retweet-edge string, e.g. '(2->1)'
    Hand-chosen based on [8]; other encodings (adjacency, incident, friendship) are mentioned but not tested.
assumptions (5)
  • domain assumption The propagation tree reconstruction algorithm in Section 3.3 produces trees close to the true retweet cascades.
    X only provides star graphs; the tree is inferred using two conditions and a random fallback (Eq. 1). If wrong, the graph information in the prompt is unreliable.
  • domain assumption Cosine similarity between embedding vectors captures the content and structure similarity relevant to astroturf coordination.
    Balanced RAG selects examples by Eq. 3; the retrievals are only as meaningful as the embeddings.
  • domain assumption The label distribution and examples in the training set are representative of the test distribution.
    Few-shot and RAG use training labels at inference; the transductive setup assumes train/test overlap in style.
  • domain assumption A frozen decoder-only LLM can reason reliably over text-encoded graph structures and few-shot examples.
    The whole framework relies on this ability; no analysis of failure modes is provided.
  • domain assumption FakeNewsNet/PolitiFact labels correctly identify coordinated astroturf campaigns.
    Ground truth is inherited from prior dataset construction without re-validation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Network-informed Prompt Engineering against Organized Astroturf Campaigns under Extreme Class Imbalance." pith.science (2026). https://pith.science/paper/O4DHV2M3

@misc{pith2026250111849,
  author       = {Pith},
  title        = {Pith review of: Network-informed Prompt Engineering against Organized Astroturf Campaigns under Extreme Class Imbalance},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/O4DHV2M3}},
  note         = {Machine review of arXiv:2501.11849}
}
read the original abstract

Detecting organized political campaigns is of paramount importance in fighting against disinformation on social media. Existing approaches for the identification of such organized actions employ techniques mostly from network science, graph machine learning and natural language processing. Their ultimate goal is to analyze the relationships and interactions (e.g. re-posting) among users and the textual similarities of their posts. Despite their effectiveness in recognizing astroturf campaigns, these methods face significant challenges, notably the class imbalance in available training datasets. To mitigate this issue, recent methods usually resort to data augmentation or increasing the number of positive samples, which may not always be feasible or sufficient in real-world settings. Following a different path, in this paper, we propose a novel framework for identifying astroturf campaigns based solely on large language models (LLMs), introducing a Balanced Retrieval-Augmented Generation (Balanced RAG) component. Our approach first gives both textual information concerning the posts (in our case tweets) and the user interactions of the social network as input to a language model. Then, through prompt engineering and the proposed Balanced RAG method, it effectively detects coordinated disinformation campaigns on X (Twitter). The proposed framework does not require any training or fine-tuning of the language model. Instead, by strategically harnessing the strengths of prompt engineering and Balanced RAG, it facilitates LLMs to overcome the effects of class imbalance and effectively identify coordinated political campaigns. The experimental results demonstrate that by incorporating the proposed prompt engineering and Balanced RAG methods, our framework outperforms the traditional graph-based baselines, achieving 2x-3x improvements in terms of precision, recall and F1 scores.

Figures

Figures reproduced from arXiv: 2501.11849 by the authors.

Figure 1
Figure 1. Overview of the proposed framework. We first construct a propagation tree [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Comparison of different metrics (precision, recall, [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 14 canonical work pages

  1. [1]

    Dimacali, Ojasmitha Pedirappagari, Namyong Park, Franck Dernoncourt, Yu Wang, Nikos Kanakaris, Hanieh Deilamsalehy, Ryan A

    Steven Au, Cameron J. Dimacali, Ojasmitha Pedirappagari, Namyong Park, Franck Dernoncourt, Yu Wang, Nikos Kanakaris, Hanieh Deilamsalehy, Ryan A. Rossi, and Nesreen K. Ahmed. 2025. Personalized Graph-Based Retrieval for Large Language Models. arXiv:2501.02157 [cs.CL] https://arxiv.org/abs/2501.02157

    arXiv 2025

  2. [2]

    Devipsita Bhattacharya and Sudha Ram. 2012. Sharing news articles using 140 characters: A diffusion analysis on Twitter. In 2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining . IEEE, 966–971

    work page 2012

  3. [3]

    Alexandre Bovet and Hernán A Makse. 2019. Influence of fake news in Twitter during the 2016 US presidential election. Nature communications 10, 1 (2019), 1–14. Network-informed Prompt Engineering against Organized Astroturf Campaigns WWW Companion ’25, April 28-May 2, 2025, Sydney, NSW, Australia

    work page 2019

  4. [4]

    Deborah Bunker. 2020. Who do you trust? The digital destruction of shared situational awareness and the COVID-19 infodemic. International Journal of Information Management 55 (2020), 102201

    work page 2020

  5. [5]

    Ahmed, Theodore Willke, and Yizhou Sun

    Kewei Cheng, Nesreen K. Ahmed, Theodore Willke, and Yizhou Sun. 2024. Structure Guided Prompt: Instructing Large Language Model in Multi-Step Rea- soning by Exploring Graph Structure of the Text. arXiv:2402.13415 [cs.CL] https://arxiv.org/abs/2402.13415

    arXiv 2024

  6. [6]

    Fatima Ezzeddine, Omran Ayoub, Silvia Giordano, Gianluca Nogara, Ihab Sbeity, Emilio Ferrara, and Luca Luceri. 2023. Exposing influence campaigns in the age of LLMs: a behavioral-based AI approach to detecting state-sponsored trolls. EPJ Data Science 12, 1 (2023), 46

    work page 2023

  7. [7]

    Ahmed, Franck Dernoncourt, Ryan A

    Jiangnan Fang, Cheng-Tse Liu, Jieun Kim, Yash Bhedaru, Ethan Liu, Nikhil Singh, Nedim Lipka, Puneet Mathur, Nesreen K. Ahmed, Franck Dernoncourt, Ryan A. Rossi, and Hanieh Deilamsalehy. 2024. Multi-LLM Text Summarization. arXiv:2412.15487 [cs.CL] https://arxiv.org/abs/2412.15487

    arXiv 2024

  8. [8]

    Bahare Fatemi, Jonathan Halcrow, and Bryan Perozzi. 2023. Talk like a graph: Encoding graphs for large language models. arXiv preprint arXiv:2310.04560 (2023)

    arXiv 2023

Show all 28 references
  1. [9]

    Miriam Fernandez and Harith Alani. 2018. Online misinformation: Challenges and future directions. In Companion Proceedings of the The Web Conference 2018 . 595–602

    2018

  2. [10]

    Deen Freelon and Tetyana Lokot. 2020. Russian Twitter disinformation campaigns reach across the American political spectrum. Misinformation Review (2020)

    2020

  3. [11]

    Louie Giray. 2023. Prompt engineering with ChatGPT: a guide for academic writers. Annals of biomedical engineering 51, 12 (2023), 2629–2633

    2023

  4. [12]

    Nir Grinberg, Kenneth Joseph, Lisa Friedland, Briony Swire-Thompson, and David Lazer. 2019. Fake news on Twitter during the 2016 US presidential election. Science 363, 6425 (2019), 374–378

    2019

  5. [13]

    Yi Han, Shanika Karunasekera, and Christopher Leckie. 2020. Graph Neural Networks with Continual Learning for Fake News Detection from Social Media. CoRR abs/2007.03316 (2020). arXiv:2007.03316

    2020 arXiv

  6. [14]

    Franziska B Keller, David Schoch, Sebastian Stier, and JungHwan Yang. 2017. How to manipulate social media: Analyzing political astroturfing using ground truth data from South Korea. In Eleventh international AAAI conference on Web and Social Media

    2017

  7. [15]

    Franziska B Keller, David Schoch, Sebastian Stier, and JungHwan Yang. 2020. Political astroturfing on twitter: How to coordinate a disinformation campaign. Political Communication 37, 2 (2020), 256–280

    2020

  8. [16]

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, et al. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing S...

    2020

  9. [17]

    Huayang Li, Yixuan Su, Deng Cai, Yan Wang, and Lemao Liu. 2022. A survey on retrieval-augmented text generation. arXiv preprint arXiv:2202.01110 (2022)

    2022 arXiv

  10. [18]

    Alexander Loth, Martin Kappes, and Marc-Oliver Pahl. 2024. Blessing or curse? A survey on the Impact of Generative AI on Fake News. arXiv:2404.03021 [cs.CL] https://arxiv.org/abs/2404.03021

    2024 arXiv

  11. [19]

    Luca Luceri, Jinyi Ye, Julie Jiang, and Emilio Ferrara. 2024. The Susceptibility Paradox in Online Social Influence. arXiv preprint arXiv:2406.11553 (2024)

    2024 arXiv

  12. [20]

    Dimitrios Michail, Nikos Kanakaris, and Iraklis Varlamis. 2022. Detection of fake news campaigns using graph convolutional networks. International Journal of Information Management Data Insights 2, 2 (2022), 100104

    2022

  13. [21]

    Shaswata Mitra, Trisha Chakraborty, Subash Neupane, Aritran Piplai, and Sudip Mittal. 2024. Use of Graph Neural Networks in Aiding Defensive Cyber Opera- tions. arXiv:2401.05680 [cs.CR] https://arxiv.org/abs/2401.05680

    2024 arXiv

  14. [22]

    Bronstein

    Federico Monti, Fabrizio Frasca, Davide Eynard, Damon Mannion, and Michael M. Bronstein. 2019. Fake News Detection on Social Media using Geometric Deep Learning. CoRR abs/1902.06673 (2019). arXiv:1902.06673

    2019 arXiv

  15. [23]

    Bryan Perozzi, Bahare Fatemi, Dustin Zelle, Anton Tsitsulin, Mehran Kazemi, Rami Al-Rfou, and Jonathan Halcrow. 2024. Let your graph do the talking: Encoding structured data for llms. arXiv preprint arXiv:2402.05862 (2024)

    2024 arXiv

  16. [24]

    David Schoch, Franziska B Keller, Sebastian Stier, and JungHwan Yang. 2022. Co- ordination patterns reveal online political astroturfing across the world.Scientific reports 12, 1 (2022), 4572

    2022

  17. [25]

    Samia Tasnim, Md Mahbub Hossain, and Hoimonty Mazumder. 2020. Impact of rumors and misinformation on COVID-19 in social media. Journal of preventive medicine and public health 53, 3 (2020), 171–174

    2020

  18. [26]

    Joanna Taylor and Claudia Pagliari. 2018. Mining social media data: How are research sponsors and researchers addressing the ethical challenges? Research Ethics 14, 2 (2018), 1–39

    2018

  19. [27]

    Iraklis Varlamis, Dimitrios Michail, Foteini Glykou, and Panagiotis Tsantilas

  20. [2022]

    Fake" coordinated political campaign or

    A survey on the use of graph convolutional networks for combating fake news. Future Internet 14, 3 (2022), 70. A Implementation details The source code is available at https://github.com/nkanak/brag- fake-news-campaigns. Due to X’s privacy policies, the utilized dataset cannot...

    2022

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.