---
title: "A Multimodal Framework for Financial Fake News Detection for"
description: "The authors build a detection framework for Brazilian Portuguese financial news. 98.34% accuracy. 19 references."
authors: ["José Vitor Souza Cardoso Requena", "João Victor Assaoka Ribeiro", "Lilian Berton"]
year: 2026
venue: "Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1"
source: "https://aclanthology.org/2026.propor-1.23/"
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/requena-2026-multimodal-framework-financial-fake-news/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/requena-2026-multimodal-framework-financial-fake-news/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/20/) / A Multimodal Framework for Financial Fake News Detection for Brazilian Portuguese

# A Multimodal Framework for Financial Fake News Detection for Brazilian Portuguese

**Authors:** [José Vitor Souza Cardoso Requena](https://md.research.elamur.ai/free/en/authors/jose-vitor-souza-cardoso-requena/), [João Victor Assaoka Ribeiro](https://md.research.elamur.ai/free/en/authors/joao-victor-assaoka-ribeiro/) and [Lilian Berton](https://md.research.elamur.ai/free/en/authors/lilian-berton/) · **Year:** 2026 · **Venue:** [Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1](https://md.research.elamur.ai/free/en/venues/propor/) ([International Conference on Computational Processing of Portuguese 2026](https://md.research.elamur.ai/free/en/venues/propor/2026/)) · **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/2026.propor-1.23/) · [PDF](https://aclanthology.org/2026.propor-1.23.pdf)

## Summary
Research on Portuguese fake news has mostly covered politics, public health and everyday topics, and nearly all detectors read plain text only. This ignores screenshots and image-based scams, which are common in finance. The authors build a detection framework for Brazilian Portuguese financial news.

They filter Fake.BR and FakeRecogna to keep only economy and finance articles. They add legitimate articles scraped from BBC News Brasil, giving a balanced set of 3434 items. A shared cleaning pipeline is applied: NFKC normalization, URL/HTML removal, and symbol and whitespace normalization. They then fine-tune BERTimbau-Base for real/fake classification and compare it with an SVM + TF-IDF baseline over five random 80/20 splits. BERTimbau averages 98.34% accuracy and beats the SVM on all reported metrics.

The "multimodal" element is an OCR step: Tesseract turns screenshots into text, which the same classifier then labels. This branch is shown in a web demo but not evaluated with numbers. The work gives a domain-specific baseline and a practical prototype for Portuguese financial fact-checking.

## Key points
- Data: finance-filtered subsets of Fake.BR (Monteiro et al., 2018) and FakeRecogna (Garcia et al., 2022), plus real news scraped from BBC News Brasil. The total is 3434 balanced documents.
- Model: BERTimbau-Base fine-tuned with 256-token inputs, batch size 8, AdamW at a 2e-5 learning rate, linear warmup and 3 epochs. No stemming or lemmatization was applied.
- Result: mean accuracy of 98.34% across 5 stratified runs with low variance, higher than SVM + TF-IDF. The authors highlight high recall as important for catching deceptive financial content.
- Multimodal branch: Tesseract OCR extracts text from screenshots such as headlines, bank alerts and messages, and the fine-tuned model then classifies it. This is demonstrated on an Agência Lupa example only.
- Tooling and limitations: the system is served through a FastAPI web service built with PyTorch, Hugging Face Transformers and NLTK. The dataset is small, the focus is Brazil-only, and the authors made no comparison with other architectures or LLMs.

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

📚 Shelves (2): [🖥️ 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

🧭 Topics (1): [🖼️ Vision & language](https://md.research.elamur.ai/free/en/topics/multimodal/) · 657 works

## Cite

Cite as: José Vitor Souza Cardoso Requena, João Victor Assaoka Ribeiro and Lilian Berton. 2026. A Multimodal Framework for Financial Fake News Detection for Brazilian Portuguese. Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1.

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

```bibtex
@inproceedings{requena2026multimodal,
  title = {{A Multimodal Framework for Financial Fake News Detection for Brazilian Portuguese}},
  author = {José Vitor Souza Cardoso Requena, João Victor Assaoka Ribeiro and Lilian Berton},
  booktitle = {Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1},
  year = {2026},
  url = {https://aclanthology.org/2026.propor-1.23/}
}
```

- [← previous A Multilingual Red Teaming–Driven Safety Analysis of LLMs](https://md.research.elamur.ai/free/en/papers/pandeiro-2026-multilingual-red-teaming-driven-safety/)
- [next → A multimodal system to support disaster risk management](https://md.research.elamur.ai/free/en/papers/lucena-2026-sistema-multimodal-de-apoio-ao/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/20/)

## References

19 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Hugo Queiroz Abonizio, Janaina Ignacio De Morais, Gabriel Marques Tavares, and Sylvio Barbon Junior. 2020. Language-independent fake news detection: English, portuguese, and spanish mutual features. Future Internet, 12(5):87. [find](https://scholar.google.com/scholar?q=Hugo+Queiroz+Abonizio%2C+Janaina+Ignacio+De+Morais%2C+Gabriel+Marques+Tavares%2C+and+Sylvio+Barbon+Junior%2E+2020%2E+Language%2Dindependent+fake+news+detection%3A+English%2C+portuguese%2C+and+spanish+mutual+features%2E+F)
2. Agência Lupa. 2025. Perfis no Facebook simulam canais do governo federal para aplicar golpe de phishing. Acesso em: 06 dez. 2025. [find](https://scholar.google.com/scholar?q=Ag%C3%AAncia+Lupa%2E+2025%2E+Perfis+no+Facebook+simulam+canais+do+governo+federal+para+aplicar+golpe+de+phishing%2E+Acesso+em%3A+06+dez%2E+2025%2E)
3. Anísio Pereira Batista Filho, Débora da Con-ceiçao Araújo, Máverick André Dionísio Ferreira, and Paulo Salgado Gomes de Mattos Neto. 2021. Fake news detection about covid-19 in the portuguese language. In Encontro Nacional de Inteligência Artificial e Computacional (ENIAC), pages 492–503. SBC. [find](https://scholar.google.com/scholar?q=An%C3%ADsio+Pereira+Batista+Filho%2C+D%C3%A9bora+da+Con%2Dcei%C3%A7ao+Ara%C3%BAjo%2C+M%C3%A1verick+Andr%C3%A9+Dion%C3%ADsio+Ferreira%2C+and+Paulo+Salgado+Gomes+de+Mattos+Neto%2E+2021%2E+Fake+news+detection+about+covid%2D19+in+the+portuguese+language)
4. Flavio Carvalho, Helder Yukio Okuno, Lais Baroni, and Gustavo Guedes. 2020. A brazilian portuguese moral foundations dictionary for fake news classification. In 2020 39th International Conference of the Chilean Computer Science Society (SCCC), pages 1–5. IEEE. [find](https://scholar.google.com/scholar?q=Flavio+Carvalho%2C+Helder+Yukio+Okuno%2C+Lais+Baroni%2C+and+Gustavo+Guedes%2E+2020%2E+A+brazilian+portuguese+moral+foundations+dictionary+for+fake+news+classification%2E+In+2020+39th+International+Conference+of+t)
5. 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,…
6. Pedro Henrique Arruda Faustini and Thiago Ferreira Covoes. 2020. Fake news detection in multiple platforms and languages. Expert Systems with Applications, 158:113503. [find](https://scholar.google.com/scholar?q=Pedro+Henrique+Arruda+Faustini+and+Thiago+Ferreira+Covoes%2E+2020%2E+Fake+news+detection+in+multiple+platforms+and+languages%2E+Expert+Systems+with+Applications%2C+158%3A113503%2E)
7. Renato M Silva, Roney LS Santos, Tiago A Almeida, and Thiago AS Pardo. 2020. Towards automatically filtering fake news in portuguese. Expert Systems with Applications, 146:113199. [find](https://scholar.google.com/scholar?q=Renato+M+Silva%2C+Roney+LS+Santos%2C+Tiago+A+Almeida%2C+and+Thiago+AS+Pardo%2E+2020%2E+Towards+automatically+filtering+fake+news+in+portuguese%2E+Expert+Systems+with+Applications%2C+146%3A113199%2E)
8. Marcelo Fischer, Rejwanul Haque, Paul Stynes, and Pramod Pathak. 2022. Identifying fake news in brazilian portuguese. In International Conference on Applications of Natural Language to Information Systems, pages 111–118. Springer. [find](https://scholar.google.com/scholar?q=Marcelo+Fischer%2C+Rejwanul+Haque%2C+Paul+Stynes%2C+and+Pramod+Pathak%2E+2022%2E+Identifying+fake+news+in+brazilian+portuguese%2E+In+International+Conference+on+Applications+of+Natural+Language+to+Information+Sys)
9. Shivangi Singhal, Tanisha Pandey, Saksham Mrig, Rajiv Ratn Shah, and Ponnurangam Kumaraguru. 2022. Leveraging intra and inter modality relationship for multimodal fake news detection. In Companion proceedings of the Web conference 2022, pages 726–734. [find](https://scholar.google.com/scholar?q=Shivangi+Singhal%2C+Tanisha+Pandey%2C+Saksham+Mrig%2C+Rajiv+Ratn+Shah%2C+and+Ponnurangam+Kumaraguru%2E+2022%2E+Leveraging+intra+and+inter+modality+relationship+for+multimodal+fake+news+detection%2E+In+Companion+pro)
10. Gabriel L Garcia, Luis CS Afonso, and João P Papa. 2022. Fakerecogna: A new brazilian corpus for fake news detection. In International Conference on Computational Processing of the Portuguese Language, pages 57–67. Springer. [find](https://scholar.google.com/scholar?q=Gabriel+L+Garcia%2C+Luis+CS+Afonso%2C+and+Jo%C3%A3o+P+Papa%2E+2022%2E+Fakerecogna%3A+A+new+brazilian+corpus+for+fake+news+detection%2E+In+International+Conference+on+Computational+Processing+of+the+Portuguese+Language)
11. Fábio Souza, Rodrigo Nogueira, and Roberto Lotufo. 2020. Bertimbau: pretrained bert models for brazilian portuguese. In Brazilian conference on intelligent systems, pages 403–417. Springer. [find](https://scholar.google.com/scholar?q=F%C3%A1bio+Souza%2C+Rodrigo+Nogueira%2C+and+Roberto+Lotufo%2E+2020%2E+Bertimbau%3A+pretrained+bert+models+for+brazilian+portuguese%2E+In+Brazilian+conference+on+intelligent+systems%2C+pages+403%E2%80%93417%2E+Springer%2E)
12. Shimon Kogan, Tobias J Moskowitz, and Marina Niessner. 2018. Fake news: Evidence from financial markets. Available at SSRN 3231461. [find](https://scholar.google.com/scholar?q=Shimon+Kogan%2C+Tobias+J+Moskowitz%2C+and+Marina+Niessner%2E+2018%2E+Fake+news%3A+Evidence+from+financial+markets%2E+Available+at+SSRN+3231461%2E)
13. Lígia Venturott and Ruslan Mitkov. 2021. Fake news detection for portuguese with deep learning. In Proceedings of the Translation and Interpreting Technology Online Conference, pages 149–153. [find](https://scholar.google.com/scholar?q=L%C3%ADgia+Venturott+and+Ruslan+Mitkov%2E+2021%2E+Fake+news+detection+for+portuguese+with+deep+learning%2E+In+Proceedings+of+the+Translation+and+Interpreting+Technology+Online+Conference%2C+pages+149%E2%80%93153%2E)
14. Rafael A Monteiro, Roney LS Santos, Thiago AS Pardo, Tiago A De Almeida, Evandro ES Ruiz, and Oto A Vale. 2018. Contributions to the study of fake news in portuguese: New corpus and automatic detection results. In International Conference on Computational Processing of the Portuguese Language, pages… [find](https://scholar.google.com/scholar?q=Rafael+A+Monteiro%2C+Roney+LS+Santos%2C+Thiago+AS+Pardo%2C+Tiago+A+De+Almeida%2C+Evandro+ES+Ruiz%2C+and+Oto+A+Vale%2E+2018%2E+Contributions+to+the+study+of+fake+news+in+portuguese%3A+New+corpus+and+automatic+detectio)
15. Maik Paixão, Rinaldo Lima, and Bernard Espinasse. 2020. Fake news classification and topic modeling in brazilian portuguese. In 2020 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT), pages 427–432. IEEE. [find](https://scholar.google.com/scholar?q=Maik+Paix%C3%A3o%2C+Rinaldo+Lima%2C+and+Bernard+Espinasse%2E+2020%2E+Fake+news+classification+and+topic+modeling+in+brazilian+portuguese%2E+In+2020+IEEE%2FWIC%2FACM+International+Joint+Conference+on+Web+Intelligence+and)
16. Anshul Vikram Pandey, Katharina Rall, Margaret L Satterthwaite, Oded Nov, and Enrico Bertini. 2015. How deceptive are deceptive visualizations? an empirical analysis of common distortion techniques. In Proceedings of the 33rd annual acm conference on human factors in computing systems, pages 1469–14… [find](https://scholar.google.com/scholar?q=Anshul+Vikram+Pandey%2C+Katharina+Rall%2C+Margaret+L+Satterthwaite%2C+Oded+Nov%2C+and+Enrico+Bertini%2E+2015%2E+How+deceptive+are+deceptive+visualizations%3F+an+empirical+analysis+of+common+distortion+techniques%2E+I)
17. Vinícius Baião Pires, Daniel Guerreiro, and 1 others. 2024. Portuguese fake news classification with bert models. In Encontro Nacional de Inteligência Artificial e Computacional (ENIAC), pages 834–845. SBC. [find](https://scholar.google.com/scholar?q=Vin%C3%ADcius+Bai%C3%A3o+Pires%2C+Daniel+Guerreiro%2C+and+1+others%2E+2024%2E+Portuguese+fake+news+classification+with+bert+models%2E+In+Encontro+Nacional+de+Intelig%C3%AAncia+Artificial+e+Computacional+%28ENIAC%29%2C+pages+834%E2%80%93845)
18. Reuters Fact Check. 2023. Checagem de fatos: Governo não autorizou auditoria para cortar 2 milhões de beneficiários do Bolsa Família. Acesso em: 05 dez. 2025. [find](https://scholar.google.com/scholar?q=Reuters+Fact+Check%2E+2023%2E+Checagem+de+fatos%3A+Governo+n%C3%A3o+autorizou+auditoria+para+cortar+2+milh%C3%B5es+de+benefici%C3%A1rios+do+Bolsa+Fam%C3%ADlia%2E+Acesso+em%3A+05+dez%2E+2025%2E)
19. João Filipe Carriço Rodrigues. 2020. Fake news classification in european portuguese language. Master’s thesis, ISCTE-Instituto Universitario de Lisboa (Portugal). [find](https://scholar.google.com/scholar?q=Jo%C3%A3o+Filipe+Carri%C3%A7o+Rodrigues%2E+2020%2E+Fake+news+classification+in+european+portuguese+language%2E+Master%E2%80%99s+thesis%2C+ISCTE%2DInstituto+Universitario+de+Lisboa+%28Portugal%29%2E)

*Updated 1 November 2026*