Quantum Enhanced Multimodal Analysis of Political Polarization on TikTok: A Case Study of Ethiopia’s Digital Public Sphere
Main Article Content
Abstract
Ethiopia's political landscape, characterized by ethnic federalism and the National Dialogue process, faces escalating polarization amplified by TikTok's algorithmically curated content. Traditional machine learning approaches struggle to capture the quantum-like dynamics of political discourse, superposition of identities, entanglement of ethnic and ideological factors, and context-dependent meaning. This study introduces the first quantum-enhanced multimodal framework for analyzing political discourse on Ethiopian TikTok, integrating quantum entanglement-driven fake news detection (Q-ALIGNer), quantum LSTM sentiment analysis, and quantum frequency-based opinion shift modeling (OpinionXf). We developed a hybrid quantum-classical pipeline processing 50,000 Amharic, Oromo, Tigrinya, and English TikTok videos. Q-ALIGNer encodes text, video, and audio modalities as quantum states with entanglement-based fusion. Quantum LSTM captures temporal sentiment evolution, while OpinionXf models opinion shifts using frequency-domain transformations. Performance was evaluated against classical baselines using 10-fold cross-validation. Q-ALIGNER achieved 92.5% accuracy, outperforming classical models by 8.2–13.9%, with only 4.6% accuracy drop under adversarial attack versus 11.9% for classical models. Quantum LSTM achieved 89.7% accuracy with 15.2% MAE reduction over AfriBERTa. Sarcasm detection improved by 8.4% and coded political language by 9.1%. OpinionXf achieved 85.7% precision and 100% recall for 72-hour early warning, detecting shifts 3–6 days before classical models. Ablation study revealed quantum layers contributed 46.3% and entanglement 53.7% of total performance gain. Entanglement-based similarity maps revealed three political actor clusters with intra-cluster entanglement 0.85–0.92 versus inter-cluster 0.65–0.72. Quantum-enhanced frameworks significantly improve detection of misinformation, sentiment polarization, and opinion shifts in Ethiopian political discourse, enabling proactive early warning systems. Deploy the 3-layer quantum model with all-to-all entanglement for Ethiopia's National Dialogue Commission, prioritizing high-persuadability local issues while approaching identity-based topics through deliberative processes.
Article Details
References
Addis Standard. (2025, November 6). Genesis of two digital cities: Social media as dual polis; rise of virtual serfdom in Ethiopia. Addis Standard. https://addisstandard.com/genesis-of-two-digital-cities-social-media-as-dual-polis-rise-of-virtual-serfdom-in-ethiopia/
Bergholm, V., Izaac, J., Schuld, M., Gogolin, C., Ahmed, S., & Killoran, N. (2022). PennyLane: Automatic differentiation of hybrid quantum-classical computations. arXiv, 1811.04968. https://doi.org/10.48550/arXiv.1811.04968
Busemeyer, J. R., & Bruza, P. D. (2012). Quantum models of cognition and decision. Cambridge University Press. https://doi.org/10.1017/CBO9780511997716
Chen, Y., Yan, K., Pan, Y., & Dong, D. (2025). QiNN-QJ: A quantum-inspired neural network with quantum jump for multimodal sentiment analysis. arXiv. https://arxiv.org/abs/2510.27091
Entrena-Serrano, C., Alves, T., Soler, M., & Berenguer, A. (2025). TikTok’s research API: Problems without explanations. arXiv, 2506.09746. https://doi.org/10.48550/arXiv.2506.09746
Ethiopian National Dialogue Commission. (2025, July 1). Youth playing a key role in advancing Ethiopia’s national dialogue process. https://ethiondc.org.et/en/youth-playing-a-key-role-in-advancing-ethiopias-national-dialogue-process/
Ethiopian National Dialogue Commission. (2025, October 23). ENDC engages Tigray media professionals in consultative forum. https://ethiondc.org.et/en/endc-engages-tigray-media-professionals-in-consultative-forum/
Goshu, B. (2025). Mitigating polarization in Ethiopia’s political system: An adversarial collaboration framework. Academia Journal of Humanities & Social Sciences, 2, 314–329. https://doi.org/10.3126/ajhss.v2i1.80459
Ho, C.-L. (2017). Quantum walk in periodic potential on a line and a model of interacting opinions. arXiv. https://arxiv.org/abs/1612.03056
IBM Quantum. (2025). Quantum approximate optimization algorithm. IBM Quantum Documentation. https://quantum.cloud.ibm.com/docs/en/tutorials/quantum-approximate-optimization-algorithm
Jia, K., Meng, F., Chen, Z., Du, M., & Liang, J. (2025). LQMF-RD: A lightweight quantum-driven multi-modal fusion framework for rumor detection. *Knowledge-Based Systems, 330*(B), 114633. https://doi.org/10.1016/j.knosys.2025.114633
Khrennikova, P., & Haven, E. (2017). Voters’ preferences in a quantum framework. In E. Haven & P. Khrennikova (Eds.), The Palgrave handbook of quantum models in social science (pp. 141–155). Palgrave Macmillan. https://doi.org/10.1057/978-1-137-49276-0_8
Maksymov, I. S., & Pogrebna, G. (2024). Quantum-mechanical modelling of asymmetric opinion polarisation in social networks. Entropy, 26(8), 653. https://doi.org/10.3390/e26080653
Nekatibeb, H. (2025). Amharic language sentiment analysis using deep learning in the case of Ethiopia Broadcasting Corporation [Master’s thesis, St. Mary’s University]. St. Mary’s University Institutional Repository. http://repository.smuc.edu.et/handle/123456789/8778
Ogueji, K., Zhu, Y., & Lin, J. (2021). Small data? No problem! Exploring the viability of pretrained multilingual language models for low-resourced languages. In Proceedings of the 1st Workshop on Multilingual Representation Learning (pp. 116–126). Association for Computational Linguistics. https://aclanthology.org/2021.mrl-1.11
PennyLane. (2025). PyTorch interface. PennyLane Documentation. https://docs.pennylane.ai/en/stable/introduction/interfaces/torch.html
Pothos, E. M., & Busemeyer, J. R. (2022). Quantum cognition. Annual Review of Psychology, 73, 749–778. https://doi.org/10.1146/annurev-psych-033020-123501
Qiskit Contributors. (2025). Running with multiple-GPUs and/or multiple nodes. Qiskit Aer Documentation. https://qiskit.github.io/qiskit-aer/howtos/running_gpu.html
Razo, L. (2025, January 24). Barcelona and Europe lead global innovation in quantum science and democracy. European City Squared (EC²) Project. https://journalofcyberpolicy.com/barcelona-and-europe-lead-global-innovation-in-quantum-science-and-democracy/
Sharma, K. P., Lal, G., & Shukla, M. (2025). Quantum behaved binary gravitational search algorithm with random forest for twitter spammer detection. Results in Engineering, 25, 103993. https://doi.org/10.1016/j.rineng.2025.103993
Sharath, P. E. (2025, July 4). Quantum-enhanced AI for stock market prediction: A multi-agent approach with QLSTM. Medium. https://medium.com/@sharath.pe/quantum-enhanced-ai-for-stock-market-prediction-a-multi-agent-approach-with-qlstm-afe16d1687af
Singh, J., Bhangu, K. S., Alkhanifer, A., AlZubi, A. A., & Ali, F. (2025). Quantum neural networks for multimodal sentiment, emotion, and sarcasm analysis. Alexandria Engineering Journal, 124, 170–187. https://doi.org/10.1016/j.aej.2025.06.088
Taye, E. (2025). Deep learning-based Amharic text classification with semantic understanding [Master’s thesis, St. Mary’s University]. St. Mary’s University Institutional Repository. http://repository.smuc.edu.et/handle/123456789/8774
Tehsin, S., Nasir, I. M., Abdelbaki, W., Alrowais, F., Abualhamayel, R., Yahya, A. E., & Marzouk, R. (2026). Q-ALIGNer: A quantum entanglement-driven multimodal framework for robust fake news detection. Computers, Materials & Continua, 87(2), 71. https://doi.org/10.32604/cmc.2026.076514
Thakur, R., Chaturvedi, H., Shah, R., Chauhan, J., & Sharma, A. (2025). Frequency-based quantum deep learning models capture opinion shifts in deliberative discourse. arXiv. https://arxiv.org/abs/2509.22603
Zaman, K., Ahmed, T., Kashif, M., Hanif, M. A., Marchisio, A., & Shafique, M. (2024). Studying the impact of quantum-specific hyperparameters on hybrid quantum-classical neural networks. arXiv, 2402.10605. https://doi.org/10.48550/arXiv.2402.10605
Zhang, Y., Li, Y., & Ren, Y. (2025). Quantum-inspired neural networks for multimodal sentiment analysis. Engineering Applications of Artificial Intelligence, 163, 117–129. https://doi.org/10.1016/j.engappai.2025.109876