REMIX-FND: A Multi-Modal Domain-Invariant Framework with Adaptive Evidence Retrieval for Cross-Domain Fake News Detection
Article Sidebar
Main Article Content
Abstract
Moreover, it generalizes well, facilitates evidence utilization, and remains applicable as generated text by machines proliferates. In contrast, existing approaches frequently release incomplete models rather than a complete, deployable, and reproducible stack.
REMIX-FND is a deployable, open-source framework that fills this existing gap. We combine multi-modal fusion with a novel application of Domain-Invariant Meta-Learning (DIML), meaning domain adversarial and MAML-style training. In addition, Monte Carlo dropout is employed for uncertainty-conditioned evidence retrieval depth, a Dynamic Source Reliability Graph (DSRG) for temporally decaying source reliability, and a six-detector ensemble for AI-generated text detection. The stack provides a reference API for tiered explanations, as well as tag-stratified diagnostics on a fixed test file.
On the standard FakeNewsNet split, domain adversarial training boosts the weighted F1 score from 85.22 to 85.47 compared to a baseline of DistilRoBERTa, but a preliminary DIML checkpoint fails to perform as well as this run on the same split. Six-class stance classification on LIAR achieves 46.25% accuracy and outperforms classical CNN anchor methods on this dataset. The authors focus on the implementation of integrated open and harness measurements. Statistical aggregation, full OOD protocols, external accuracy of AI-text benchmarks, and latency analysis are left to future work.
Downloads
Article Details
Section
This work is licensed under a Creative Commons Attribution 4.0 International License.
All articles published in Interdisciplinary Journal of AI, Machine Learning & Data Science (IJAIMLDS) are licensed under the terms of the Creative Commons Attribution 4.0 International License (CC BY 4.0).
This license allows others to share, copy, distribute, and adapt the work, provided that proper credit is given to the original author(s) and the source.
Authors retain copyright and grant Interdisciplinary Journal of AI, Machine Learning & Data Science (IJAIMLDS) the right of first publication.
How to Cite
References
1. Lazer, D.M.J., Baum, M.A., Benkler, Y., et al.: The science of fake news. Science 359(6380), 1094–1096 (2018). https://doi.org/10.1126/science.aao2998 DOI: https://doi.org/10.1126/science.aao2998
2. World Health Organization: Managing the COVID-19 infodemic. WHO Technical Report (2020). https://www.who.int/news-room/feature-stories/detail/managing-the-covid-19-infodemic-promoting-healthy-behaviours-and-mitigating-the-harm-from-misinformation-and-disinformation
3. Zhou, X., Xue, Y., Zafarani, R.: Out-of-distribution evidence-aware fake news detection via dual adversarial debiasing. In: Proc. AAAI, vol. 37, no. 4, pp. 5596–5604 (2023). https://scholar.google.com/scholar?q=Out-of-distribution+evidence-aware+fake+news+detection+via+dual+adversarial+debiasing+Zhou+Xue+Zafarani+AAAI+2023
4. Liu, Z., Xiong, C., Sun, M., Liu, Z.: MUSER: A multi-step evidence retrieval enhancement framework. In: Proc. ACL, pp. 2791–2803 (2022). https://aclanthology.org/2022.acl-long.194/
5. Chen, Y., Mao, Y., Zhu, H., Ying, Y.: Unsupervised domain-agnostic fake news detection using multi-modal weak signals. In: Proc. ACM MM, pp. 5537–5546 (2022). https://doi.org/10.1145/3503161.3548113 DOI: https://doi.org/10.1145/3503161.3548113
6. Sadasivan, V.S., Kumar, A., Balasubramanian, S., Wang, W., Feizi, S.: Can AI-generated text be reliably detected? Stress testing AI text detectors under various attacks. Trans. Mach. Learn. Res. (2025). https://openreview.net/forum?id=OOgsAZdFOt
7. Zhang, X., Cao, J., Li, X., Sheng, Q., Zhong, L., Shu, K.: Mining dual emotion for fake news detection. In: Proc. WWW, pp. 3465–3476 (2021). https://doi.org/10.1145/3442381.3450004 DOI: https://doi.org/10.1145/3442381.3450004
8. Qi, P., Cao, J., Yang, T., Guo, J., Li, J.: Exploiting multi-domain visual information for fake news detection. In: Proc. ICDM, pp. 518–527 (2019). https://doi.org/10.1109/ICDM.2019.00062 DOI: https://doi.org/10.1109/ICDM.2019.00062
9. Wang, Y., Ma, F., Jin, Z., et al.: EANN: Event adversarial neural networks for multi-modal fake news detection. In: Proc. KDD, pp. 849–857 (2018). https://doi.org/10.1145/3219819.3219903 DOI: https://doi.org/10.1145/3219819.3219903
10. Finn, C., Abbeel, P., Levine, S.: Model-agnostic meta-learning for fast adaptation. In: Proc. ICML, pp. 1126–1135 (2017). https://proceedings.mlr.press/v70/finn17a.html
11. Ganin, Y., Ustinova, E., Ajakan, H., et al.: Domain-adversarial training of neural networks. J. Mach. Learn. Res. 17(59), 1–35 (2016). https://www.jmlr.org/papers/v17/15-239.html
12. Gal, Y., Ghahramani, Z.: Dropout as a Bayesian approximation: Representing model uncertainty in deep learning. In: Proc. ICML, pp. 1050–1059 (2016). https://proceedings.mlr.press/v48/gal16.html
13. Guo, C., Pleiss, G., Sun, Y., Weinberger, K.Q.: On calibration of modern neural networks. In: Proc. ICML, pp. 1321–1330 (2017). https://proceedings.mlr.press/v70/guo17a.html
14. Wang, W.Y.: “Liar, liar pants on fire”: A new benchmark dataset for fake news detection. In: Proc. ACL, pp. 422–426 (2017). https://aclanthology.org/P17-2067/ DOI: https://doi.org/10.18653/v1/P17-2067
15. Shu, K., Mahudeswaran, D., Wang, S., Lee, D., Liu, H.: FakeNewsNet: A data repository with news content, social context, and spatiotemporal information for studying fake news on social media. Big Data 8(3), 171–188 (2020). https://doi.org/10.1089/big.2020.0062 DOI: https://doi.org/10.1089/big.2020.0062
16. Guo, B., Zhang, X., Zhang, Z., et al.: How close is ChatGPT to human experts? Comparison corpus, evaluation, and detection. arXiv:2301.07597 (2023, preprint). https://arxiv.org/abs/2301.07597
17. McNemar, Q.: Note on the sampling error of the difference between correlated proportions or percentages. Psychometrika 12(2), 153–157 (1947). https://doi.org/10.1007/BF02295996 DOI: https://doi.org/10.1007/BF02295996