---
title: "MemeMind at ArAIEval Shared Task: Spotting Persuasive Spans"
description: "The work offers a simple, reproducible baseline for span-level persuasion detection in a morphologically… 22 references, cited by 1 paper here."
authors: ["Md. Rafiul Biswas", "Zubair Shah", "Wajdi Zaghouani"]
year: 2024
venue: "Proceedings of the Second Arabic Natural Language Processing Conference"
source: "https://aclanthology.org/2024.arabicnlp-1.50/"
doi: "10.18653/v1/2024.arabicnlp-1.50"
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/biswas-2024-mememind-araieval-shared-task-spotting/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/biswas-2024-mememind-araieval-shared-task-spotting/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2024](https://md.research.elamur.ai/free/en/papers/2024/page/15/) / MemeMind at ArAIEval Shared Task: Spotting Persuasive Spans in Arabic Text with Persuasion Techniques Identification

# MemeMind at ArAIEval Shared Task: Spotting Persuasive Spans in Arabic Text with Persuasion Techniques Identification

**Authors:** [Md. Rafiul Biswas](https://md.research.elamur.ai/free/en/authors/md-rafiul-biswas/), [Zubair Shah](https://md.research.elamur.ai/free/en/authors/zubair-shah/) and [Wajdi Zaghouani](https://md.research.elamur.ai/free/en/authors/wajdi-zaghouani/) · **Year:** 2024 · **Venue:** [Proceedings of the Second Arabic Natural Language Processing Conference](https://md.research.elamur.ai/free/en/venues/arabicnlp/) ([ArabicNLP 2024](https://md.research.elamur.ai/free/en/venues/arabicnlp/2024/)) · **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/2024.arabicnlp-1.50/) · [PDF](https://aclanthology.org/2024.arabicnlp-1.50.pdf) · [DOI 10.18653/v1/2024.arabicnlp-1.50](https://doi.org/10.18653/v1/2024.arabicnlp-1.50) · [OpenAlex](https://openalex.org/W4402671253)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · [1](https://md.research.elamur.ai/free/en/papers/biswas-2024-mememind-araieval-shared-task-spotting/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
The authors tackle ArAIEval 2024 Task 1, which asks systems to find which stretches of Arabic tweets and news paragraphs use persuasion or propaganda techniques and to name the technique for each stretch. They reframe the problem as token-level sequence labelling. The character offsets of each labelled span become BIO tags per technique, and an AraBERT-base encoder with a linear-plus-softmax head assigns a tag to every subword. Training runs in two stages. First the pre-trained encoder is frozen and only the head is fitted. Then all weights are unfrozen and updated jointly. Among several Arabic and multilingual BERT variants, AraBERTv2 gave the best development scores. The submitted system reached a micro-F1 of 0.2774 on the test set, which placed third on the leaderboard. The very low macro-F1 values show that rare techniques remain largely undetected. The work offers a simple, reproducible baseline for span-level persuasion detection in a morphologically rich, dialect-heavy language with little annotated data.

## Key points
- The ArAIEval Task 1 data has 6,997 training, 921 validation and 1,046 test snippets annotated with 23 techniques. The labels are heavily skewed: "Loaded Language" accounts for 55.69% of spans, "Name Calling Labeling" for 14.23% and "Exaggeration-Minimisation" for 6.83%.
- Fine-tuning ran in two phases with batch size 32:
  - 10 epochs training only the head, at learning rate 2e-5.
  - 5 epochs of full fine-tuning, at learning rate 1e-5.
  - Training took about 6 hours on one RTX 2080 Ti.
- On the development set, AraBERTv2 scored micro-F1 0.277 and macro-F1 0.058. The other models scored lower on micro-F1:
  - Original AraBERT: 0.237.
  - CAMeLBERT-da: 0.199.
  - A multilingual model reported as mBERT in the table (named MARBERT in the text): 0.130.
- On the official test set the system reached precision 0.2446 and recall 0.3202, for a micro-F1 of 0.2774. This ranked third, behind CUET_sstm (0.2995) and Mela (0.2833); a random baseline scored 0.0151.
- In error analysis, context-dependent techniques such as "Whataboutism" were frequently missed. The model also often confused "Loaded Language" with "Name Calling/Labeling", since both depend on emotionally charged wording.

## Interesting ideas
- The encoder is frozen first while a fresh head learns to map stable pretrained features onto labels, and only then is the whole network unfrozen at a lower rate; this keeps the noisy early gradients of a random head from damaging pretrained weights.
- Character-offset span annotations are converted into per-technique BIO tags on subword tokens, which reduces overlapping span detection to ordinary sequence tagging.
- Micro-F1 and macro-F1 are reported side by side, and the large gap between them shows that a good-looking headline score can hide near-zero performance on rare classes.
- The error analysis groups mistakes by label pairs that share surface cues, such as pejorative words that fit two techniques, rather than just listing per-class scores.
- After the deadline, adding part-of-speech and named-entity features helped separate techniques that look alike (reported only as unofficial results).

## For Mova

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

- **I1** (tree): gparse could warm up new arc/label heads on a frozen transformer before full fine-tuning, which is useful when adding a new UD treebank or dialect with few sentences.
- **I2** (frames): spans of speech acts or irony cues in prag could be stored as offsets and projected onto tokens via BIO, making every flagged token traceable back to its annotated span.
- **I3** (learn): evaluation reports for deprels, frame roles or speech-act classes should always pair micro and macro scores, so that skewed labels cannot mask failures on rare categories.
- **I4** (frames): a confusion review of semantically adjacent prag categories (e.g. sarcasm vs. plain negative evaluation) could point to where explicit, explainable disambiguation rules are needed.
- **I5** (frames): pragmatic classifiers in prag could draw on the UD tree and entity types already produced upstream, instead of relying on surface words alone.

*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

- [Hamad bin Khalifa University](https://md.research.elamur.ai/free/en/universities/hamad-bin-khalifa-university/) — University, Doha, Qatar · 144 works · 14,175 citations
- [Qatar Foundation](https://md.research.elamur.ai/free/en/ngos/qatar-foundation/) — Nonprofit, Doha, Qatar · 17 works · 2,111 citations
- [Northwestern University in Qatar](https://md.research.elamur.ai/free/en/universities/northwestern-university-in-qatar/) — University, Doha, Qatar · 39 works · 148 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,954 works · [🤝 Sociology and Political Science](https://md.research.elamur.ai/free/en/shelves/social-sciences/sociology-and-political-science/) · 626 works

🧭 Topics (1): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works

## Cite

Cite as: Md. Rafiul Biswas, Zubair Shah and Wajdi Zaghouani. 2024. MemeMind at ArAIEval Shared Task: Spotting Persuasive Spans in Arabic Text with Persuasion Techniques Identification. Proceedings of the Second Arabic Natural Language Processing Conference.

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

```bibtex
@inproceedings{biswas2024mememind,
  title = {{MemeMind at ArAIEval Shared Task: Spotting Persuasive Spans in Arabic Text with Persuasion Techniques Identification}},
  author = {Md. Rafiul Biswas, Zubair Shah and Wajdi Zaghouani},
  booktitle = {Proceedings of the Second Arabic Natural Language Processing Conference},
  year = {2024},
  doi = {10.18653/v1/2024.arabicnlp-1.50},
  url = {https://doi.org/10.18653/v1/2024.arabicnlp-1.50}
}
```

- [← previous LomonosovMSU at SemEval-2024 Task 4: Comparing LLMs and embedder models to identifying propaganda techniques in the content of memes in English for subtasks No1, No2a, and No2b](https://md.research.elamur.ai/free/en/papers/skiba-2024-lomonosovmsu-semeval-2024-task-4/)
- [next → Mhm... Yeah? Okay! Evaluating the Naturalness and Communicative Function of Synthesized Feedback Responses in Spoken Dialogue](https://md.research.elamur.ai/free/en/papers/figueroa-2024-mhm-yeah-okay-evaluating-naturalness/)
- [↑ back to papers of 2024](https://md.research.elamur.ai/free/en/papers/2024/page/15/)

## References

22 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Malak Abdullah, Ola Altiti, and Rasha Obiedat. 2022. Detecting Propaganda Techniques in English News Articles using Pre-trained Transformers. In 2022 13th International Conference on Information and Communication Systems (ICICS), pages 301–308. IEEE. [find](https://scholar.google.com/scholar?q=Malak+Abdullah%2C+Ola+Altiti%2C+and+Rasha+Obiedat%2E+2022%2E+Detecting+Propaganda+Techniques+in+English+News+Articles+using+Pre%2Dtrained+Transformers%2E+In+2022+13th+International+Conference+on+Information+and+C)
2. Firoj Alam, Abul Hasnat, Fatema Ahmed, Md Arid Hasan, and Maram Hasanain. 2024. [ArMeme: Propagandistic Content in Arabic Memes](https://md.research.elamur.ai/free/en/papers/alam-2024-armeme-propagandistic-content-arabic-memes/). arXiv preprint arXiv:2406.03916. [↗](https://arxiv.org/abs/2406.03916)
3. Firoj Alam, Shaden Shaar, Fahim Dalvi, Hassan Sajjad, Alex Nikolov, Hamdy Mubarak, Giovanni Da San Martino, Ahmed Abdelali, Nadir Durrani, Kareem Darwish, et al. 2021. Fighting the COVID-19 infodemic: modeling the perspective of journalists, fact-checkers, social media platforms, policy makers, and the society. In Findings of EMNLP-2021. [↗](https://doi.org/10.48550/arxiv.2005.00033)
4. Wissam Antoun, Fady Baly, and Hazem Hajj. 2020. [AraBERT: Transformer-based model for arabic language understanding](https://md.research.elamur.ai/free/en/papers/antoun-2020-arabert-transformer-based-model-arabic/). In Proceedings of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools, with a Shared Task on Offensive Language Detection, pages 9–15. [↗](https://doi.org/10.48550/arxiv.2003.00104)
5. Alberto Barrón-Cedeño, Giovanni Da San Martino, Israa Jaradat, and Preslav Nakov. 2019. PropPy: A System to Unmask Propaganda in Online News. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 72–77. [find](https://scholar.google.com/scholar?q=Alberto+Barr%C3%B3n%2DCede%C3%B1o%2C+Giovanni+Da+San+Martino%2C+Israa+Jaradat%2C+and+Preslav+Nakov%2E+2019%2E+PropPy%3A+A+System+to+Unmask+Propaganda+in+Online+News%2E+In+Proceedings+of+the+57th+Annual+Meeting+of+the+Associati)
6. Anis Charfi, Wajdi Zaghouani, Sayed Hamid Mehdi, and Esraa Mohamed. 2019. [A fine-grained annotated multi-dialectal Arabic corpus](https://md.research.elamur.ai/free/en/papers/charfi-2019-fine-grained-annotated-multi-dialectal/). In Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2019), pages 198–204.
7. In Proceedings of the First Arabic Natural Language Processing Conference (ArabicNLP 2023). Maram Hasanain, Md. Arid Hasan, Fatema Ahmed, Reem Suwaileh, Md. [find](https://scholar.google.com/scholar?q=In+Proceedings+of+the+First+Arabic+Natural+Language+Processing+Conference+%28ArabicNLP+2023%29%2E+Maram+Hasanain%2C+Md%2E+Arid+Hasan%2C+Fatema+Ahmed%2C+Reem+Suwaileh%2C+Md%2E)
8. Rafiul Biswas, Wajdi Zaghouani, and Firoj Alam. 2024b. [ArAIEval Shared Task: Propagandistic Techniques Detection in Unimodal and Multimodal Arabic Content](https://md.research.elamur.ai/free/en/papers/hasanain-2024-araieval-shared-task-propagandistic-techniques/). In Proceedings of the Second Arabic Natural Language Processing Conference (ArabicNLP 2024), Bangkok. Association for Computational Linguistics.
9. Akib Mohi Ud Din Khanday, Qamar Rayees Khan, and Syed Tanzeel Rabani. 2021. Detecting textual propaganda using machine learning techniques. Baghdad Science Journal, 18(1):0199–0199. [↗](https://doi.org/10.21123/bsj.2021.18.1.0199)
10. Giovanni Da San Martino, Seunghak Yu, Alberto Barrón-Cedeño, Rostislav Petrov, and Preslav Nakov. 2019. [Fine-grained analysis of propaganda in news articles](https://md.research.elamur.ai/free/en/papers/martino-2019-fine-grained-analysis-propaganda-news/). In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 5636–5646. [↗](https://doi.org/10.48550/arxiv.1910.02517)
11. Jon Dehdari, Lamia Tounsi, and Josef van Genabith. 2011. [Morphological features for parsing morphologically-rich languages: A case of arabic](https://md.research.elamur.ai/free/en/papers/dehdari-2011-morphological-features-parsing-morphologically-rich/). In Proceedings of the Second Workshop on Statistical Parsing of Morphologically Rich Languages, pages 12–21.
12. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. [BERT: Pre-training of deep bidirectional transformers for language understanding](https://md.research.elamur.ai/free/en/papers/devlin-2019-bert-pre-training-deep-bidirectional/). In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 4171–4186. [↗](https://doi.org/10.18653/v1/n19-1423)
13. Jinfen Li and Lu Xiao. 2020. syrapropa at semeval-2020 task 11: Bert-based models design for propagandistic technique and span detection. arXiv preprint arXiv:2008.10163. [↗](https://arxiv.org/abs/2008.10163)
14. Nora Mapes, Isaiah Onando Mulang, Maynard Jude Samonte, Jaime Rafael Angel Stein, Tobias Kreutz, Henning Wachsmuth, and Matthias Hagen. 2019. Divisive Language and Propaganda Detection using Multi-head Attention Transformers. In Proceedings of the Second Workshop on Natural Language Processing for Internet Freedom: Censorship, Disinformation, and Propaganda, pages 80–89. [find](https://scholar.google.com/scholar?q=Nora+Mapes%2C+Isaiah+Onando+Mulang%2C+Maynard+Jude+Samonte%2C+Jaime+Rafael+Angel+Stein%2C+Tobias+Kreutz%2C+Henning+Wachsmuth%2C+and+Matthias+Hagen%2E+2019%2E+Divisive+Language+and+Propaganda+Detection+using+Multi%2Dhea)
15. Giovanni Da San Martino, Stefano Cresci, Alberto Barrón-Cedeño, Seunghak Yu, Roberto Di Pietro, and Preslav Nakov. 2020. A survey on computational propaganda detection. arXiv preprint arXiv:2007.08024. [↗](https://arxiv.org/abs/2007.08024)
16. Hajder S Rabiee. 2011. [Adapting standard open-source resources to tagging a morphologically rich language: a case study with arabic](https://md.research.elamur.ai/free/en/papers/rabiee-2011-adapting-standard-open-source-resources/). In Proceedings of the Second Student Research Workshop associated with RANLP 2011, pages 127–132.
17. Ibrahim Abu Farha, Wajdi Zaghouani, and Walid Magdy. 2021. [Overview of the WANLP 2021 shared task on sarcasm and sentiment detection in Arabic](https://md.research.elamur.ai/free/en/papers/farha-2021-overview-wanlp-2021-shared-task/). In Proceedings of the Sixth Arabic Natural Language Processing Workshop, pages 296–305.
18. Chantal Shurafa, Kareem Darwish, and Wajdi Zaghouani. 2020. Political framing: US COVID19 blame game. In Social Informatics: 12th International Conference, SocInfo 2020, Pisa, Italy, October 6–9, 2020, Proceedings 12, pages 333–351. [find](https://scholar.google.com/scholar?q=Chantal+Shurafa%2C+Kareem+Darwish%2C+and+Wajdi+Zaghouani%2E+2020%2E+Political+framing%3A+US+COVID19+blame+game%2E+In+Social+Informatics%3A+12th+International+Conference%2C+SocInfo+2020%2C+Pisa%2C+Italy%2C+October+6%E2%80%939%2C+2020)
19. Maram Hasanain, Fatema Ahmed, and Firoj Alam. 2023a. Large language models for propaganda span annotation. arXiv preprint arXiv:2311.09812. [↗](https://arxiv.org/abs/2311.09812)
20. 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](https://md.research.elamur.ai/free/en/papers/veeramani-2023-knowtellconvince-araieval-shared-task-disinformation/). In Proceedings of ArabicNLP 2023, pages 519–524.
21. Maram Hasanain, Fatema Ahmed, and Firoj Alam. 2024a. Can gpt-4 identify propaganda? annotation and detection of propaganda spans in news articles. In Proceedings of the 2024 Joint International Conference On Computational Linguistics, Language Resources And Evaluation, LREC-COLING 2024, Torino, Italy. [find](https://scholar.google.com/scholar?q=Maram+Hasanain%2C+Fatema+Ahmed%2C+and+Firoj+Alam%2E+2024a%2E+Can+gpt%2D4+identify+propaganda%3F+annotation+and+detection+of+propaganda+spans+in+news+articles%2E+In+Proceedings+of+the+2024+Joint+International+Confer)
22. Maram Hasanain, Firoj Alam, Hamdy Mubarak, Samir Abdaljalil, Wajdi Zaghouani, Preslav Nakov, Giovanni Da San Martino, and Abed Alhakim Freihat. 2023b. [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/).

*Updated 1 November 2026*