---
title: "KnowTellConvince at ArAIEval Shared Task: Disinformation and"
description: "The authors describe their entry to the ArAIEval 2023 shared task, which targets two problems in… 24 references, cited by 4 papers here."
authors: ["Hariram Veeramani", "Surendrabikram Thapa", "Usman Naseem"]
year: 2023
venue: "Proceedings of ArabicNLP 2023"
source: "https://aclanthology.org/2023.arabicnlp-1.49/"
doi: "10.18653/v1/2023.arabicnlp-1.49"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/veeramani-2023-knowtellconvince-araieval-shared-task-disinformation/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/veeramani-2023-knowtellconvince-araieval-shared-task-disinformation/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2023](https://md.research.elamur.ai/free/en/papers/2023/page/12/) / KnowTellConvince at ArAIEval Shared Task: Disinformation and Persuasion Detection in Arabic using Similar and Contrastive Representation Alignment

# KnowTellConvince at ArAIEval Shared Task: Disinformation and Persuasion Detection in Arabic using Similar and Contrastive Representation Alignment

**Authors:** [Hariram Veeramani](https://md.research.elamur.ai/free/en/authors/hariram-veeramani/), [Surendrabikram Thapa](https://md.research.elamur.ai/free/en/authors/surendrabikram-thapa/) and [Usman Naseem](https://md.research.elamur.ai/free/en/authors/usman-naseem/) · **Year:** 2023 · **Venue:** [Proceedings of ArabicNLP 2023](https://md.research.elamur.ai/free/en/venues/arabicnlp/) ([ArabicNLP 2023](https://md.research.elamur.ai/free/en/venues/arabicnlp/2023/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2023.arabicnlp-1.49/) · [PDF](https://aclanthology.org/2023.arabicnlp-1.49.pdf) · [DOI 10.18653/v1/2023.arabicnlp-1.49](https://doi.org/10.18653/v1/2023.arabicnlp-1.49) · [OpenAlex](https://openalex.org/W4389518427)

**Citations:** 3 citations (OpenAlex, 2026-10-14) · [4](https://md.research.elamur.ai/free/en/papers/veeramani-2023-knowtellconvince-araieval-shared-task-disinformation/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
The authors describe their entry to the ArAIEval 2023 shared task, which targets two problems in Arabic text where far fewer tools exist than for English. Subtask 1A is a yes/no decision on whether a tweet or news paragraph uses persuasion (propaganda) techniques. Task 2 looks at tweets: 2A flags disinformation, and 2B sorts it into hate speech, offensive content, rumor or spam.

Their system ensembles four classifiers, all built on BERT Arabic Base:
- one trained with a supervised contrastive objective;
- one fine-tuned directly on the task with cross-entropy;
- one first adapted on XNLI with relabeled classes and then fine-tuned on the task;
- one that turns sentence embeddings into cosine-similarity scores against positive and negative example sentences, passed through GELU.

An ablation compares baselines, pairs of components and the full ensemble. The full combination wins on both binary tasks, gaining about 4 F1 points on persuasion and about 3 on disinformation over plain Arabic-BERT. The work matters as a practical recipe for low-resource propaganda and disinformation detection, though the gains come at the cost of compute and interpretability.

## Key points
- **Backbone and baselines.** Every component uses BERT Arabic Base (Safaya et al., 2020). Baselines are Arabic-BERT and a sentiment-tuned ArSAS BERT checkpoint.
- **Subtask 1A results.** Micro-F1 rises from 72.23 (Arabic-BERT) and 73.4 (ArSAS BERT) to 76.14 for Model (A+B+C+D). Macro-F1 reaches 76.82.
- **Subtask 2A results.** Micro-F1 goes from 86.4 to 89.67 with the full ensemble, and macro-F1 reaches 89.93. Model (A+D) is the strongest pair at 89.38.
- **Subtask 2B results.** Only B and B+D are reported, at 80.36 and 80.71 micro-F1. This is a small gain.
- **Class balance shift in Task 1.** The development set is roughly 78% non-propaganda and 22% propaganda. The test set is about 34.2% versus 65.8%, so the class balance nearly flips.

## Interesting ideas
- An existing NLI corpus (XNLI) is repurposed as an intermediate training step by merging entailment and neutral into one class and keeping contradiction as the other, which gives a cheap binary pretraining signal before the real task.
- One component scores an input by how close it is to positive versus negative reference sentences, so its decision can be traced to concrete exemplars rather than opaque weights.
- The ablation reports baselines, every useful pair of components and the full ensemble, which shows which parts actually add value (A+D beats A+B here).
- The persuasion data has very different class proportions in the development and test sets, a reminder that a threshold tuned on one split may be miscalibrated on another.
- A contrastive loss pulls same-label examples together and pushes different-label examples apart in embedding space, which sharpens class boundaries when data is scarce.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (learn): Mova could relabel public NLI data into "compatible vs. contradicts" pairs to pre-train a checker that tests whether a new sentence clashes with the current world of the text.
- **I2** (frames): speech-act or irony decisions in prag could fall back on nearest labeled exemplar frames and report "closest to example X" as part of the explanation.
- **I3** (learn): when Mova combines witness parsers or repair rules, it could publish pairwise and leave-one-out scores to show that each witness earns its place.
- **I4** (learn): evaluation runs could log label priors per split and include a prior-shifted negative control, so that gains are not artifacts of class balance.
- **I5** (tree): gparse could add a contrastive term over arc or label representations to better separate confusable dependency relations in small treebanks.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [University of California, Los Angeles](https://md.research.elamur.ai/free/en/universities/university-of-california-los-angeles/) — University, Los Angeles, United States · 87 works · 54,961 citations
- [Virginia Tech](https://md.research.elamur.ai/free/en/universities/virginia-tech/) — University, Blacksburg, United States · 36 works · 43,036 citations
- [James Cook University](https://md.research.elamur.ai/free/en/universities/james-cook-university/) — University, Townsville, Australia · 3 works · 23 citations

📚 Shelves (5): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,772 works · [🗄️ Information Systems](https://md.research.elamur.ai/free/en/shelves/computer-science/information-systems/) · 1,437 works · [🤝 Sociology and Political Science](https://md.research.elamur.ai/free/en/shelves/social-sciences/sociology-and-political-science/) · 577 works

🧭 Topics (1): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 2,908 works

## Cite

Cite as: Hariram Veeramani, Surendrabikram Thapa and Usman Naseem. 2023. KnowTellConvince at ArAIEval Shared Task: Disinformation and Persuasion Detection in Arabic using Similar and Contrastive Representation Alignment. Proceedings of ArabicNLP 2023.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{veeramani2023knowtellconvince,
  title = {{KnowTellConvince at ArAIEval Shared Task: Disinformation and Persuasion Detection in Arabic using Similar and Contrastive Representation Alignment}},
  author = {Hariram Veeramani, Surendrabikram Thapa and Usman Naseem},
  booktitle = {Proceedings of ArabicNLP 2023},
  year = {2023},
  doi = {10.18653/v1/2023.arabicnlp-1.49},
  url = {https://doi.org/10.18653/v1/2023.arabicnlp-1.49}
}
```

- [← previous ivrit.ai: A Comprehensive Dataset of Hebrew Speech for AI Research and Development](https://md.research.elamur.ai/free/en/papers/marmor-2023-ivrit-ai-comprehensive-dataset-hebrew/)
- [next → LANS: Large-scale Arabic News Summarization Corpus](https://md.research.elamur.ai/free/en/papers/alhamadani-2023-lans-large-scale-arabic-news/)
- [↑ back to papers of 2023](https://md.research.elamur.ai/free/en/papers/2023/page/12/)

## References

24 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Khaled M Fouad, Sahar F Sabbeh, and Walaa Medhat. 2022. Arabic fake news detection using deep learning. Computers, Materials & Continua, 71(2). [↗](https://doi.org/10.32604/cmc.2022.021449)
2. Areej Al-Hassan and Hmood Al-Dossari. 2022. Detection of hate speech in arabic tweets using deep learning. Multimedia systems, 28(6):1963–1974. [find](https://scholar.google.com/scholar?q=Areej+Al%2DHassan+and+Hmood+Al%2DDossari%2E+2022%2E+Detection+of+hate+speech+in+arabic+tweets+using+deep+learning%2E+Multimedia+systems%2C+28%286%29%3A1963%E2%80%931974%2E)
3. Firoj Alam, Stefano Cresci, Tanmoy Chakraborty, Fabrizio Silvestri, Dimiter Dimitrov, Giovanni Da San Martino, Shaden Shaar, Hamed Firooz, and Preslav Nakov. 2022a. A survey on multimodal disinformation detection. In Proceedings of the 29th International Conference on Computational Linguistics, pages 6625–6643. [↗](https://doi.org/10.48550/arxiv.2103.12541)
4. Firoj Alam, Hamdy Mubarak, Wajdi Zaghouani, Giovanni Da San Martino, and Preslav Nakov. 2022b. [Overview of the wanlp 2022 shared task on propaganda detection in arabic](https://md.research.elamur.ai/free/en/papers/alam-2022-overview-wanlp-2022-shared-task/). WANLP 2022, page 108.
5. Nuha Albadi, Maram Kurdi, and Shivakant Mishra. 2018. Are they our brothers? analysis and detection of religious hate speech in the arabic twittersphere. In 2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), pages 69–76. IEEE. [find](https://scholar.google.com/scholar?q=Nuha+Albadi%2C+Maram+Kurdi%2C+and+Shivakant+Mishra%2E+2018%2E+Are+they+our+brothers%3F+analysis+and+detection+of+religious+hate+speech+in+the+arabic+twittersphere%2E+In+2018+IEEE%2FACM+International+Conference+on+A)
6. Abdullah M Alkadri, Abeer Elkorany, and Cherry Ahmed. 2022. Enhancing detection of arabic social spam using data augmentation and machine learning. Applied Sciences, 12(22):11388. [↗](https://doi.org/10.3390/app122211388)
7. Suchin Gururangan, Ana Marasović, Swabha Swayamdipta, Kyle Lo, Iz Beltagy, Doug Downey, and Noah A Smith. 2020. Don’t stop pretraining: Adapt language models to domains and tasks. arXiv preprint arXiv:2004.10964. [↗](https://arxiv.org/abs/2004.10964)
8. Ivan Habernal, Raffael Hannemann, Christian Pollak, Christopher Klamm, Patrick Pauli, and Iryna Gurevych. 2017. Argotario: Computational argumentation meets serious games. EMNLP 2017, page 7. [find](https://scholar.google.com/scholar?q=Ivan+Habernal%2C+Raffael+Hannemann%2C+Christian+Pollak%2C+Christopher+Klamm%2C+Patrick+Pauli%2C+and+Iryna+Gurevych%2E+2017%2E+Argotario%3A+Computational+argumentation+meets+serious+games%2E+EMNLP+2017%2C+page+7%2E)
9. Ivan Habernal, Patrick Pauli, and Iryna Gurevych. 2018. Adapting serious game for fallacious argumentation to german: Pitfalls, insights, and best practices. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018). [find](https://scholar.google.com/scholar?q=Ivan+Habernal%2C+Patrick+Pauli%2C+and+Iryna+Gurevych%2E+2018%2E+Adapting+serious+game+for+fallacious+argumentation+to+german%3A+Pitfalls%2C+insights%2C+and+best+practices%2E+In+Proceedings+of+the+Eleventh+Internation)
10. Fouzi Harrag and Mohamed Khalil Djahli. 2022. Arabic fake news detection: A fact checking based deep learning approach. Transactions on Asian and Low-Resource Language Information Processing, 21(4):1– 34. [↗](https://doi.org/10.1145/3501401)
11. Maram Hasanain, Firoj Alam, Hamdy Mubarak, Samir Abdaljalil, Wajdi Zaghouani, Preslav Nakov, Giovanni Da San Martino, and Abed Alhakim Freihat. 2023. [ArAIEval Shared Task: Persuasion Techniques and Disinformation Detection in Arabic Text](https://md.research.elamur.ai/free/en/papers/hasanain-2023-araieval-shared-task-persuasion-techniques/). In Proceedings of the First Arabic Natural Language Processing Conference (ArabicNLP 2023), Singapore. Association for Computational Linguistics.
12. Sanaa Kaddoura, Suja A Alex, Maher Itani, Safaa Henno, Asma AlNashash, and D Jude Hemanth. 2023. Arabic spam tweets classification using deep learning. Neural Computing and Applications, pages 1–14. [↗](https://doi.org/10.1007/s00521-023-08614-w)
13. Rajaraman Kanagasabai, Saravanan Rajamanickam, Hariram Veeramani, Adam Westerski, and Kim Jung Jae. 2023. Semantists at ImageArg-2023: Exploring Cross-modal Contrastive and Ensemble Models for Multimodal Stance and Persuasiveness Classification. In Proceedings of the 10th Workshop on Argument Mining, Online and in Singapore. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Rajaraman+Kanagasabai%2C+Saravanan+Rajamanickam%2C+Hariram+Veeramani%2C+Adam+Westerski%2C+and+Kim+Jung+Jae%2E+2023%2E+Semantists+at+ImageArg%2D2023%3A+Exploring+Cross%2Dmodal+Contrastive+and+Ensemble+Models+for+Multimo)
14. Preslav Nakov, Firoj Alam, Shaden Shaar, Giovanni Da San Martino, and Yifan Zhang. 2021. [A second pandemic? analysis of fake news about covid-19 vaccines in qatar](https://md.research.elamur.ai/free/en/papers/nakov-2021-second-pandemic-analysis-fake-news/). In Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021), pages 1010–1021.
15. Preslav Nakov, Alberto Barrón-Cedeño, Giovanni Da San Martino, Firoj Alam, Julia Maria Struß, Thomas Mandl, Rubén Míguez, Tommaso Caselli, Mucahid Kutlu, Wajdi Zaghouani, et al. 2022. The clef-2022 checkthat! lab on fighting the covid-19 infodemic and fake news detection. In European Conference on Information Retrieval, pages 416–428. Springer. [find](https://scholar.google.com/scholar?q=Preslav+Nakov%2C+Alberto+Barr%C3%B3n%2DCede%C3%B1o%2C+Giovanni+Da+San+Martino%2C+Firoj+Alam%2C+Julia+Maria+Stru%C3%9F%2C+Thomas+Mandl%2C+Rub%C3%A9n+M%C3%ADguez%2C+Tommaso+Caselli%2C+Mucahid+Kutlu%2C+Wajdi+Zaghouani%2C+et+al%2E+2022%2E+The+clef%2D2022+ch)
16. Ali Safaya, Moutasem Abdullatif, and Deniz Yuret. 2020. [KUISAIL at SemEval-2020 task 12: BERT-CNN for offensive speech identification in social media](https://md.research.elamur.ai/free/en/papers/safaya-2020-kuisail-semeval-2020-task-12/). In Proceedings of the Fourteenth Workshop on Semantic Evaluation, pages 2054–2059, Barcelona (online). International Committee for Computational Linguistics. In Proceedings of the second Workshop on Information Extraction from Scientific Publications, Online. Association for Computational Linguistics.
17. Hariram Veeramani, Surendrabikram Thapa, and Usman Naseem. 2023b. [DialectNLU at NADI 2023 Shared Task: Transformer Based Multitask Approach Jointly Integrating Dialect and Machine Translation Tasks in Arabic](https://md.research.elamur.ai/free/en/papers/veeramani-2023-dialectnlu-nadi-2023-shared-task/). In Proceedings of The First Arabic Natural Language Processing Conference (ArabicNLP 2023), Singapore. Association for Computational Linguistics.
18. Hariram Veeramani, Surendrabikram Thapa, and Usman Naseem. 2023c. Enhancing ESG Impact Type Identification through Early Fusion and Multilingual Models. In Proceedings of the Sixth Workshop on Financial Technology and Natural Language Processing (FinNLP). Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Hariram+Veeramani%2C+Surendrabikram+Thapa%2C+and+Usman+Naseem%2E+2023c%2E+Enhancing+ESG+Impact+Type+Identification+through+Early+Fusion+and+Multilingual+Models%2E+In+Proceedings+of+the+Sixth+Workshop+on+Financi)
19. Hariram Veeramani, Surendrabikram Thapa, and Usman Naseem. 2023d. [LowResContextQA at Qur’an QA 2023 Shared Task: Temporal and Sequential Representation Augmented Question Answering Span Detection in Arabic](https://md.research.elamur.ai/free/en/papers/veeramani-2023-lowrescontextqa-qur-qa-2023-shared/). In Proceedings of The First Arabic Natural Language Processing Conference (ArabicNLP 2023), Singapore. Association for Computational Linguistics.
20. Hariram Veeramani, Surendrabikram Thapa, and Usman Naseem. 2023e. LowResourceNLU at BLP-2023 Task 1 2: Enhancing Sentiment Classification and Violence Incitement Detection in Bangla Through Aggregated Language Models. In Proceedings of the 1st International Workshop on Bangla Language Processing (BLP-2023), Singapore. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Hariram+Veeramani%2C+Surendrabikram+Thapa%2C+and+Usman+Naseem%2E+2023e%2E+LowResourceNLU+at+BLP%2D2023+Task+1+2%3A+Enhancing+Sentiment+Classification+and+Violence+Incitement+Detection+in+Bangla+Through+Aggregated)
21. Hariram Veeramani, Surendrabikram Thapa, and Usman Naseem. 2023f. Temporally Dynamic Session-Keyword Aware Sequential Recommendation system. In 2023 International Conference on Data Mining Workshops (ICDMW). [find](https://scholar.google.com/scholar?q=Hariram+Veeramani%2C+Surendrabikram+Thapa%2C+and+Usman+Naseem%2E+2023f%2E+Temporally+Dynamic+Session%2DKeyword+Aware+Sequential+Recommendation+system%2E+In+2023+International+Conference+on+Data+Mining+Workshops+%28)
22. Mohamad Sharara, Wissam Mohamad, Ralph Tawil, Ralph Chobok, Wolf Assi, and Antonio Tannoury. 2022. [Arabert model for propaganda detection](https://md.research.elamur.ai/free/en/papers/sharara-2022-arabert-model-propaganda-detection/). In Proceedings of the The Seventh Arabic Natural Language Processing Workshop (WANLP), pages 520– 523.
23. Zien Sheikh Ali, Watheq Mansour, Fatima Haouari, Maram Hasanain, Tamer Elsayed, and Abdulaziz Al-Ali. 2023. Tahaqqaq: A real-time system for assisting twitter users in arabic claim verification. In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 3170–3174. [find](https://scholar.google.com/scholar?q=Zien+Sheikh+Ali%2C+Watheq+Mansour%2C+Fatima+Haouari%2C+Maram+Hasanain%2C+Tamer+Elsayed%2C+and+Abdulaziz+Al%2DAli%2E+2023%2E+Tahaqqaq%3A+A+real%2Dtime+system+for+assisting+twitter+users+in+arabic+claim+verification%2E+In+Pr)
24. Hariram Veeramani, Surendrabikram Thapa, and Usman Naseem. 2023a. Automated Citation Function Classification and Context Extraction in Astrophysics: Leveraging Paraphrasing and Question Answering. [find](https://scholar.google.com/scholar?q=Hariram+Veeramani%2C+Surendrabikram+Thapa%2C+and+Usman+Naseem%2E+2023a%2E+Automated+Citation+Function+Classification+and+Context+Extraction+in+Astrophysics%3A+Leveraging+Paraphrasing+and+Question+Answering%2E)

*Updated 1 November 2026*