Elamur Research⌕
☰
home / papers / 2026 / Signals Are Not States: Neuro-Symbolic Safeguards for Culturally Aware Classroom AI

Signals Are Not States: Neuro-Symbolic Safeguards for Culturally Aware Classroom AI

Paper · 2026 · Bagheri Nezhad, Sina

Summary

Classroom AI tools increasingly infer engagement, confusion, collaboration, or teaching quality from video, audio, and transcripts. In multicultural and multilingual classrooms, this can turn culturally specific behavior into stereotyped judgments. A quiet student may be labeled disengaged, averted gaze read as inattention, or code-switching taken as weak language skill. The author frames these overreaching inferences as a cross-cultural safety issue and proposes NSCR (Neuro-Symbolic Classroom Reasoning). The framework first turns raw multimodal input and contextual metadata into typed facts that carry confidence, provenance, and cultural scope. It then reasons over these facts with executable rules and applies policy constraints, including abstention and human review. The paper reports no experiments. Its contribution is methodological: a taxonomy of stereotype-prone inferences, a benchmark agenda, and metrics such as stereotype leakage and evidence faithfulness. This matters because, compared with end-to-end fusion models or long-context prompting, it makes classroom judgments auditable and localizes ambiguity in places that can be inspected.

Key points

  • NSCR works in four stages: perceptual grounding, symbolic abstraction into typed facts, executable reasoning, and governance (uncertainty thresholds, stereotype-risk policies, privacy rules, abstention).
  • It keeps three representation levels apart: observable facts (e.g., speaking turns), construct hypotheses (e.g., a confusion candidate), and stereotype-risk claims (e.g., low ability), which are blocked unless supported by evidence.
  • The proposed benchmark has six task families:
    • culture-conditioned state inference
    • evidence-grounded claim verification
    • multilingual/code-switched reasoning
    • cross-cultural collaboration analysis
    • counterfactual cultural robustness
    • culture-conditioned red-teaming
  • Proposed metrics include unsupported attribution rate, cultural calibration gap, counterfactual consistency, selective risk/coverage, and WER-conditioned performance. Abstention quality is treated as a primary safety metric.
  • The work appeared at StereACuLT 2026 (ACL workshop, pp. 88–100). It also recommends documenting annotator backgrounds and collecting multiple perspectives, including local educators, for high-stakes labels.

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

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksSocial Sciences · 1,954 worksInformation Systems · 1,730 worksSafety Research · 626 works

Cite

Cite as Bagheri Nezhad, Sina. 2026. Signals Are Not States: Neuro-Symbolic Safeguards for Culturally Aware Classroom AI. Proceedings of the 1st Workshop on Stereotypes Across Cultures in Language Technologies (StereACuLT 2026).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.stereacult-1.9

@inproceedings{sina2026signals,
  title = {{Signals Are Not States: Neuro-Symbolic Safeguards for Culturally Aware Classroom AI}},
  author = {Bagheri Nezhad, Sina},
  booktitle = {Proceedings of the 1st Workshop on Stereotypes Across Cultures in Language Technologies (StereACuLT 2026)},
  year = {2026},
  doi = {10.18653/v1/2026.stereacult-1.9},
  url = {https://doi.org/10.18653/v1/2026.stereacult-1.9}
}

References

41 references, 13 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Ameeta Agrawal, Andy Dang, Sina Bagheri Nezhad, Rhitabrat Pokharel, and Russell Scheinberg. 2024. Evaluating multilingual long-context models for retrieval and reasoning. In Proceedings of the Fourth Workshop on Multilingual Representation Learning (MRL 2024), pages 216–231, Miami, Florida, USA. Ass…
  2. Karan Ahuja, Dohyun Kim, Franceska Xhakaj, Virag Varga, Anne Xie, Stanley Zhang, Jay Eric Townsend, Chris Harrison, Amy Ogan, and Yuvraj Agarwal. 2019. Edusense: Practical classroom sensing at scale. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol., 3(3). find
  3. Arkar Min Aung, Anand Ramakrishnan, and Jacob R Whitehill. 2018. Who are they looking at? automatic eye gaze following for classroom observation video analysis. International Educational Data Mining Society. find
  4. Ethical deployment. Educational datasets involving human subjects may require IRB review, consent procedures, retention limits, and careful subgroup analysis. Reviewers and deployers should inspect not only model performance but also the policies encoded in the system, including anonymization and re… find
  5. Sina Bagheri Nezhad and Ameeta Agrawal. 2025. Enhancing large language models with neurosymbolic reasoning for multilingual tasks. In Proceedings of The 19th International Conference on Neurosymbolic Learning and Reasoning, volume 284 of Proceedings of Machine Learning Research, pages 1059–1076. PML… find
  6. Sina Bagheri Nezhad, Yao Li, and Ameeta Agrawal. 2026. SymCode: A neurosymbolic approach to mathematical reasoning via verifiable code generation. In Findings of the Association for Computational Linguistics: EACL 2026, pages 1489–1503, Rabat, Morocco. Association for Computational Linguistics. \#\# C… find
  7. We presented NSCR, a culturally grounded neuro-symbolic framework for mitigating stereotyped reasoning in classroom AI. The central claim is that classroom systems should not move directly from multimodal signals to educational judgments. Instead, they should separate observable evidence from constr… find
  8. Paulo Blikstein and Marcelo Worsley. 2016. Multimodal learning analytics and education data mining: using computational technologies to measure complex learning tasks. Journal of Learning Analytics, 3(2):220–238. find
  9. Su Lin Blodgett, Solon Barocas, Hal Daumé III, and Hanna Wallach. 2020. Language (technology) is power: A critical survey of “bias” in NLP. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5454– 5476, Online. Association for Computational Linguistics.
  10. Hervé Bredin, Ruiqing Yin, Juan Manuel Coria, Gregory Gelly, Pavel Korshunov, Marvin Lavechin, Diego Fustes, Hadrien Titeux, Wassim Bouaziz, and Marie-Philippe Gill. 2020. Pyannote.audio: Neural building blocks for speaker diarization. In ICASSP 2020 - 2020 IEEE International Conference on Acoustics… find
  11. Simon Buckingham Shum, Rebecca Ferguson, and Roberto Martinez-Maldonado. 2019. Human-centred learning analytics. Journal of Learning Analytics, 6(2):1–9. find
  12. Daniel Hershcovich, Stella Frank, Heather Lent, Miryam de Lhoneux, Mostafa Abdou, Stephanie Brandl, Emanuele Bugliarello, Laura Cabello Piqueras, Ilias Chalkidis, Ruixiang Cui, Constanza Fierro, Katerina Margatina, Phillip Rust, and Anders Søgaard. 2022. Challenges and strategies in cross-cultural N…
  13. Simon Buckingham Shum, Roberto Martínez-Maldonado, Yannis Dimitriadis, and Patricia Santos. 2024. Human-centred learning analytics: 2019–24. British Journal of Educational Technology, 55(3):755–768. find
  14. Mutlu Cukurova, Michail Giannakos, and Roberto Martinez-Maldonado. 2020. The promise and challenges of multimodal learning analytics. British Journal of Educational Technology, 51(5):1441–1449. find
  15. Dirk Hovy and Shannon L. Spruit. 2016. The social impact of natural language processing. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 591–598, Berlin, Germany. Association for Computational Linguistics.
  16. Daniele Di Mitri, Jan Schneider, Marcus Specht, and Hendrik Drachsler. 2018. From signals to knowledge: A conceptual model for multimodal learning analytics. Journal of Computer Assisted Learning, 34(4):338–349. find
  17. Pang Wei Koh, Shiori Sagawa, Henrik Marklund, Sang Michael Xie, Marvin Zhang, Akshay Balsubramani, Weihua Hu, Michihiro Yasunaga, Richard Lanas Phillips, Irena Gao, Tony Lee, Etienne David, Ian Stavness, Wei Guo, Berton Earnshaw, Imran Haque, Sara M Beery, Jure Leskovec, Anshul Kundaje, and 4 others… find
  18. Meng Fang, Shilong Deng, Yudi Zhang, Zijing Shi, Ling Chen, Mykola Pechenizkiy, and Jun Wang. 2024. Large language models are neurosymbolic reasoners. In Proceedings of the Thirty-Eighth AAAI Conference on Artificial Intelligence and Thirty-Sixth Conference on Innovative Applications of Artificial I… find
  19. Balaji Lakshminarayanan, Alexander Pritzel, and Charles Blundell. 2017. Simple and scalable predictive uncertainty estimation using deep ensembles. In Advances in Neural Information Processing Systems, volume 30. Curran Associates, Inc. find
  20. Luyu Gao, Aman Madaan, Shuyan Zhou, Uri Alon, Pengfei Liu, Yiming Yang, Jamie Callan, and Graham Neubig. 2023. PAL: Program-aided language models. In Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, pages 10764–10799. PMLR. find
  21. Nelson F. Liu, Kevin Lin, John Hewitt, Ashwin Paranjape, Michele Bevilacqua, Fabio Petroni, and Percy Liang. 2024. Lost in the middle: How language models use long contexts. Transactions of the Association for Computational Linguistics, 12:157–173.
  22. Yonatan Geifman and Ran El-Yaniv. 2019. SelectiveNet: A deep neural network with an integrated reject option. In Proceedings of the 36th International Conference on Machine Learning, volume 97 of Proceedings of Machine Learning Research, pages 2151–2159. PMLR. find
  23. Yun Long, Haifeng Luo, and Yu Zhang. 2024. Evaluating large language models in analysing classroom dialogue. npj Science of Learning, 9(1):60. find
  24. Jared D. T. Guerrero-Sosa, Francisco P. Romero, Víctor H. Menéndez-Domínguez, Jesus Serrano-Guerrero, Andres Montoro-Montarroso, and Jose A. Olivas. 2025. A comprehensive review of multimodal analysis in education. Applied Sciences, 15(11). find
  25. Su Mu, Meng Cui, and Xiaodi Huang. 2020. Multimodal data fusion in learning analytics: A systematic review. Sensors, 20(23). find
  26. Moin Nadeem, Anna Bethke, and Siva Reddy. 2021. StereoSet: Measuring stereotypical bias in pretrained language models. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long…
  27. Chuan Guo, Geoff Pleiss, Yu Sun, and Kilian Q. Weinberger. 2017. On calibration of modern neural networks. In Proceedings of the 34th International Conference on Machine Learning, volume 70 of Proceedings of Machine Learning Research, pages 1321– 1330. PMLR. find
  28. Nikita Nangia, Clara Vania, Rasika Bhalerao, and Samuel R. Bowman. 2020. CrowS-pairs: A challenge dataset for measuring social biases in masked language models. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Dan Hendrycks and Kevin Gimpel. 2017. A baseline for detecti…
  29. Processing (EMNLP), pages 1953–1967, Online. Association for Computational Linguistics. videos in the classroom. IEEE Transactions on Affective Computing, 14(2):1012–1027. find
  30. Sina Bagheri Nezhad and Ameeta Agrawal. 2026. Robust long-context multilingual retrieval and reasoning enabled by combined neural and symbolic techniques. Neurosymbolic Artificial Intelligence, 2:29498732261443192. find
  31. Yao-Hung Hubert Tsai, Shaojie Bai, Paul Pu Liang, J. Zico Kolter, Louis-Philippe Morency, and Ruslan Salakhutdinov. 2019. Multimodal transformer for unaligned multimodal language sequences. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 6558– 6569,…
  32. Xavier Ochoa and Marcelo Worsley. 2016. Editorial: Augmenting learning analytics with multimodal sensory data. Journal of Learning Analytics, 3(2):213– 219. find
  33. Jiani Wang, Kamil Hankour, Yuqi Zhang, Jennifer LoCasale-Crouch, and Jacob Whitehill. 2025. Classroom observation evaluation with large language models. In Proceedings of the Innovation and Responsibility in AI-Supported Education Workshop, volume 273 of Proceedings of Machine Learning Research, pag… find
  34. Theo Olausson, Alex Gu, Ben Lipkin, Cedegao Zhang, Armando Solar-Lezama, Joshua Tenenbaum, and Roger Levy. 2023. LINC: A neurosymbolic approach for logical reasoning by combining language models with first-order logic provers. In Proceedings of the 2023 Conference on Empirical Methods in Natural Lan…
  35. Marcelo Worsley, Dor Abrahamson, Paulo Blikstein, Shuchi Grover, Bertrand Schneider, and Mike Tis-senbaum. 2016. Situating multimodal learning analytics. In 12th International Conference of the Learning Sciences: Transforming Learning, Empowering Learners, ICLS 2016, pages 1346–1349. International S… find
  36. Abelardo Pardo and George Siemens. 2014. Ethical and privacy principles for learning analytics. British Journal of Educational Technology, 45(3):438–450. find
  37. Alicia Parrish, Angelica Chen, Nikita Nangia, Vishakh Padmakumar, Jason Phang, Jana Thompson, Phu Mon Htut, and Samuel Bowman. 2022. BBQ: A hand-built bias benchmark for question answering. In Findings of the Association for Computational Linguistics: ACL 2022, pages 2086–2105, Dublin, Ireland. Asso…
  38. Amir Zadeh, Minghai Chen, Soujanya Poria, Erik Cambria, and Louis-Philippe Morency. 2017. Tensor fusion network for multimodal sentiment analysis. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 1103–1114, Copenhagen, Denmark. Association for Computat…
  39. Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine Mcleavey, and Ilya Sutskever. 2023. Robust speech recognition via large-scale weak supervision. In Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, pages 28492–2…
  40. Ömer Sümer, Peter Gerjets, Ulrich Trautwein, and Enkelejda Kasneci. 2020. Automated anonymisation of visual and audio data in classroom studies. Preprint, arXiv:2001.05080. — [arXiv:2001.05080](https://arxiv.org/abs/2001.05080) find
  41. A Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2016. "why should i trust you?": Explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’16, page 1135–1144, New York, NY, USA. Association for…

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/nezhad-2026-signals-are-not-states-neuro/