Elamur Research⌕
☰
home / papers / 2026 / Cross-Cultural Transfer of Emoji Semantics and Sentiment in Financial Social Media

Cross-Cultural Transfer of Emoji Semantics and Sentiment in Financial Social Media

Paper · 2026 · Ahmed Mahrous and Roberto Di Pietro

Summary

The paper asks whether emojis in financial social media send the same sentiment signal regardless of language, platform, or asset type. If they do, sentiment models could move between communities without retraining. The authors gathered over 100M finance-related posts from 2007–2023, drawn from Twitter (EN, ES, JA, TR) and English-only StockTwits (stocks and crypto). StockTwits posts carry native bullish/bearish tags, while tweet labels were generated with GPT-5 and spot-checked by humans. The analysis compares communities on three layers: how often emojis are used, what they mean (XLM-RoBERTa centroids aligned with Procrustes), and their polarity. It then runs zero-shot transfer tests with emoji-only, text-only, and combined inputs. Usage frequencies diverge, most strongly between languages. Meaning and polarity, however, stay broadly consistent. Transfer across asset types loses little accuracy, and transfer across languages is the hardest case. Adding emojis narrows transfer gaps compared with text alone. The practical takeaway is that emojis offer compact, largely language-neutral cues for financial NLP in settings where labeled data is scarce.

Key points

  • Corpus: about 21M English, 0.65M Japanese, 0.25M Spanish and 0.23M Turkish tweets, plus 60M stock, 31M crypto and 4.7M Bitcoin StockTwits posts. A sample of 2,700 tweets was checked by hand to validate the GPT-5 labels.
  • Frequency comparison: uses JSD, Total Variation, Bhattacharyya Coefficient and Rank-Biased Overlap over the top 100 emojis. EN–JA appears to be the most divergent pair (JSD ≈ 0.51), while Stocks–Crypto is the closest (JSD ≈ 0.28).
  • Semantic alignment: mean cosine similarity is high everywhere (≈0.94–0.97). Cross-language NN@1 is very low (≤0.09), which points to stable global meaning but shifting local neighbourhoods.
  • Polarity: assessed with Spearman correlation, mean absolute difference and flip rate. Only a handful of context-dependent emojis reverse their sentiment between communities.
  • Transfer setup: TF-IDF + Logistic Regression, XLM-RoBERTa and ByT5 are tested across 27 settings, with DeBERTa-v3-base (ABSA) reported in the appendix. Data and code are released on Zenodo.

Note: the numeric values in the third and fourth bullets were read from a garbled table extraction and should be checked against the original PDF.

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

🎓King Abdullah University of Science and TechnologyJeddah · Saudi Arabia27 works · 1,084 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksDecision Sciences · 209 worksManagement Science and Operations Research · 108 worksHuman-Computer Interaction · 141 works

🧭 Topics 1Research subjects and languages the work is about.Sentiment & emotion · 972 works

Cite

Cite as Ahmed Mahrous and Roberto Di Pietro. 2026. Cross-Cultural Transfer of Emoji Semantics and Sentiment in Financial Social Media. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.1106

@inproceedings{mahrous2026cross,
  title = {{Cross-Cultural Transfer of Emoji Semantics and Sentiment in Financial Social Media}},
  author = {Ahmed Mahrous and Roberto {Di Pietro}},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1106},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1106}
}

References

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

  1. alaix14. 2019. Bitcoin tweets 2016-01-01 to 2019-03-29. https://www.kaggle.com/datasets/alaix14/bitcoin-tweets-20160101-to-20190329. Accessed 2026-04-22. ↗
  2. Francesco Audrino, Fabio Sigrist, and Daniele Balli-nari. 2020. The impact of sentiment and attention measures on stock market volatility. International Journal of Forecasting, 36(1):334–357. find
  3. Qiyu Bai, Qiong Dan, Zhe Mu, and Min Yang. 2019. A systematic review of emoji: Current research and future perspectives. Frontiers in Psychology, 10:476737. find
  4. Francesco Barbieri, Luis Espinosa Anke, and Jose Camacho-Collados. 2022. Xlm-t: Multilingual language models in twitter for sentiment analysis and beyond. In Proceedings of the thirteenth language resources and evaluation conference, pages 258–266.
  5. Francesco Barbieri, Jose Camacho-Collados, Francesco Ronzano, Luis Espinosa-Anke, Miguel Ballesteros, Valerio Basile, Viviana Patti, and Horacio Saggion. 2018. Semeval-2018 task 2: Multilingual emoji prediction. In Proceedings of The 12th International Workshop on Semantic Evaluation, pages 24–33.
  6. Francesco Barbieri, Germán Kruszewski, Francesco Ronzano, and Horacio Saggion. 2016. How cosmopolitan are emojis? exploring emoji usage and meaning over different languages with distributional semantics. In Proceedings of the 24th ACM International Conference on Multimedia, pages 531–535. find
  7. Hector E. Carrillo, Robert Pennington, and Yang Zhang. 2022. Is an emoji worth a thousand words? the effect of emoji usage on nonprofessional investors’ perceptions. Journal of Information Systems, 36(2):1–15. find
  8. Shuo Chen and Fan Xing. 2023. Understanding emojis for financial sentiment analysis. In Proceedings of the International Conference on Information Systems (ICIS). find
  9. Yu Chen, Xinyi Yang, Henry Howman, and Ruth Filik. 2024. Individual differences in emoji comprehension: Gender, age, and culture. PLOS ONE, 19(2):e0297379. find
  10. Giuseppe Colavito, Filippo Lanubile, and Nicole Novielli. 2025. Benchmarking large language models for automated labeling: The case of issue report classification. Information and Software Technology, 184:107758. find
  11. Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2020. Unsupervised cross-lingual representation learning at scale. In Proceedings of the 58th Annual Meeting of the Association fo… find
  12. J. Anthony Cookson and Marina Niessner. 2020. Why don’t we agree? evidence from a social network of investors. Journal of Finance, 75(1):173–228. find
  13. Flavio Di Palo, Prateek Singhi, and Bilal Fadlallah. 2024. Performance-guided llm knowledge distillation for efficient text classification at scale. arXiv preprint arXiv:2411.05045. find
  14. Marc-Aurèle Divernois and Damir Filipović. 2024. Stocktwits classified sentiment and stock returns. Digital Finance, 6(2):249–281. find
  15. Emojipedia. 2025. Emoji statistics. https://emojipedia.org/stats. Accessed 2026-04-22. ↗
  16. European Parliament and the Council of the European Union. 2016. Regulaton (eu) 2016/679 (general date protection regulation). Official Journal of the European Union, pages L119/1–88. find
  17. Fabrizio Gilardi, Meysam Alizadeh, and Meher Kubli. 2023. Chatgpt outperforms crowd workers for text annotation tasks. Proceedings of the National Academy of Sciences.
  18. Google Research. 2024. Mediapipe language detector. Accessed 2026-04-22. find
  19. Sharath Chandra Guntuku, Meng Li, Louis Tay, and Lyle H Ungar. 2019. Studying cultural differences in emoji usage across the east and the west. In Proceedings of the International AAAI Conference on Web and Social Media, volume 13, pages 226–235. find
  20. Sultan A. A. Hakami, Robert J. Hendley, and Paul Smith. 2021. Arabic emoji sentiment lexicon (arab-esl): A comparison between arabic and european emoji sentiment lexicons. In Proceedings of the Sixth Arabic Natural Language Processing Workshop, pages 60–71.
  21. Xingwei He, Zhenghao Lin, Yeyun Gong, A-Long Jin, Hang Zhang, Chen Lin, Jian Jiao, Siu Ming Yiu, Nan Duan, and Weizhu Chen. 2023. Annollm: Making large language models to be better crowdsourced annotators. arXiv preprint arXiv:2303.16854.
  22. Armand Joulin, Edouard Grave, Piotr Bojanowski, and Tomas Mikolov. 2017. Bag of tricks for efficient text classification. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics (EACL), pages 427–431. find
  23. Thomas K. Koch, Pablo Romero, and Clemens Stachl. 2022. Age and gender in language, emoji, and emoticon usage in instant messages. Computers in Human Behavior, 126:106990. find
  24. Nikola Ljubešić and Darja Fišer. 2016. A global analysis of emoji usage. In Proceedings of the 10th Web as Corpus Workshop (WAC-X) and the EmpiriST Shared Task, pages 82–89. Association for Computational Linguistics. find
  25. Federico Lopez. 2023. langdetect: Lightweight language detection in python. https://github.com/fede-lopez77/langdetect. Accessed 2026-04-22. ↗
  26. Xiaohui Lu, Wei Ai, Xiang Liu, Qian Li, Ning Wang, Guobin Huang, and Qiaozhu Mei. 2016. Learning from the ubiquitous language: an empirical analysis of emoji usage of smartphone users. In Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing, pages 770–780. find
  27. Marco Lui and Timothy Baldwin. 2012. langid.py: An off-the-shelf language identification tool. In Proceedings of the ACL 2012 System Demonstrations, pages 25–30.
  28. Niloofar Mahmoudi, Łukasz Piotr Olech, and Peter Docherty. 2022. A comprehensive study of domain-specific emoji meanings in sentiment classification. Computational Management Science, 19(2):159–197. find
  29. Ahmed Mahrous and Roberto Di Pietro. 2026. Replication package for ”cross-cultural transfer of emoji semantics and sentiment in financial social media” (findings of acl 2026).
  30. Ahmed Mahrous, Jens Schneider, and Roberto Di Pietro. 2023. The role of emojis in sentiment analysis of financial microblogs. In 2023 Fourth International Conference on Intelligent Data Science Technologies and Applications (IDSTA), pages 76–84. IEEE. find
  31. Hannah Miller, Daniel Kluver, Jacob Thebault-Spieker, Loren Terveen, and Brent Hecht. 2017. Understanding emoji ambiguity in context: The role of text in emoji-related miscommunication. In Proceedings of the International AAAI Conference on Web and Social Media, volume 11, pages 152–161. find
  32. Hannah Miller, Jacob Thebault-Spieker, Shuo Chang, Isaac Johnson, Loren Terveen, and Brent Hecht. 2016. “blissfully happy” or “ready to fight”: Varying interpretations of emoji. In Proceedings of the International AAAI Conference on Web and Social Media, volume 10, pages 259–268. find
  33. Adi Hidayat Nasution and Aytuğ Onan. 2024. Chatgpt label: Comparing the quality of human-generated and llm-generated annotations in low-resource language nlp tasks. IEEE Access, 12:71876–71900. find
  34. Nuno Oliveira, Paulo Cortez, and Nelson Areal. 2013. On the predictability of stock market behavior using stocktwits sentiment and posting volume. In Progress in Artificial Intelligence (EPIA 2013), volume 8154 of Lecture Notes in Computer Science, pages 355–365. Springer. find
  35. OpenAI. 2025. Gpt-5 system card. Aug 13, 2025 version. find
  36. Andrea Peñas Martínez. 2023. Bitcoin twitter sentiment dataset (2013–2023). https://www.kaggle.com/datasets/andreapenasmartinez/bitcoin-twitter-sentiment-dataset-20132023. Accessed 2026-04-22. ↗
  37. Thomas Renault. 2020. Sentiment analysis and machine learning in finance: A comparison of methods and models on one million messages. Digital Finance, 2(1):1–13. find
  38. Fabian Reschke and Johannes Oliver Strych. 2024. Emojis and stock returns. Review of Behavioral Finance, 16(2):223–233. find
  39. Agnese Sampietro, Stefan Felder, and Beat Siebenhaar. 2022. Do you kiss when you text? cross-cultural differences in the use of the kissing emojis in three whatsapp corpora. Intercultural Pragmatics, 19(2):183–208. find
  40. Jiahui Sun, Sean Lasser, and Seung Kyu Lee. 2023. Understanding emojis: Cultural influences in interpretation and choice of emojis. Journal of International and Intercultural Communication, 16(3):242–261. find
  41. L. J. Togans, Thomas Holtgraves, Giho Kwon, and T. E. M. Zelaya. 2021. Digitally saving face: An experimental investigation of cross-cultural differences in the use of emoticons and emoji. Journal of Pragmatics, 186:277–288. find
  42. Takumi Tomihira, Akihiko Otsuka, Akira Yamashita, and Takashi Satoh. 2020. Multilingual emoji prediction using bert for sentiment analysis. International Journal of Web Information Systems, 16(3):265–280. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/mahrous-2026-cross-cultural-transfer-emoji-semantics/