Elamur Research⌕
☰
home / papers / 2026 / IndustryAssetEQA: A Neurosymbolic Operational Intelligence System for Embodied Question Answering in Industrial Asset Maintenance

IndustryAssetEQA: A Neurosymbolic Operational Intelligence System for Embodied Question Answering in Industrial Asset Maintenance

Paper · 2026 · Chathurangi Shyalika, Dhaval C Patel and Amit Sheth

Summary

LLM-based maintenance assistants often give fluent but generic answers. These answers are poorly tied to sensor data, lack traceable evidence, and offer counterfactual or action advice that cannot be checked. The authors treat maintenance QA as an embodied decision task covering perception, reasoning, prediction and decision, rather than as plain text generation. Their system, IndustryAssetEQA, has several parts:

  • a fact extractor that turns telemetry windows around failures or healthy periods into structured, provenance-tagged episode records;
  • an SQLite episodic store;
  • an FMEA-KG built from ISO-based failure specifications;
  • a parametric risk estimator that simulates the effect of interventions;
  • a verifier and safety gate that check LLM outputs.

The authors tested the system on four datasets covering rotating machinery, turbofan engines, hydraulic systems and cyber-physical production systems. Against LLM-only baselines, it gave more structurally valid, better grounded and more consistent counterfactual answers. Experts also found far fewer severe overclaims. This matters for safety-critical industrial settings, where auditability and risk-anchored recommendations are needed before operators can trust AI support.

Key points

  • QA taxonomy with five task types: descriptive, temporal, diagnostic, counterfactual and action-oriented. All are episode-centric, meaning each query concerns one asset and one time window.
  • FMEA-KG was built with the EMPWR platform and queried via rdflib. It holds 63 failure modes across 9 asset categories, linked by relations such as affects, component_of, indicated_by and mitigated_by.
  • Gains over LLM-only baselines:
    • structural validity: up to +0.51
    • counterfactual accuracy: up to +0.47
    • explanation entailment: +0.64
  • Blinded expert study (10 experts, 22 QA pairs):
    • answerable responses: 97% vs 46% for the baseline
    • data grounding score: 4.5 vs 3.0 (p < 0.001)
    • severe overclaims: 2% vs 28%
    • inter-annotator agreement: Fleiss' κ = 0.63
  • Code, data and the FMEA-KG are released within IBM's AssetOpsBench. A pilot integration into enterprise agents (e.g., SAP Joule Agents) is planned. The counterfactual module is a surrogate estimator, not a full structural causal model.

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

🏢IBMArmonk · United States139 works · 70,691 citations🎓University of South CarolinaColumbia · United States8 works · 785 citations

 Fields, topics and concepts

📚 Shelves 3The field of science this work belongs to (OpenAlex fields).Engineering · 133 worksControl and Systems Engineering · 45 worksSafety, Risk, Reliability and Quality · 10 works

🧭 Topics 1Research subjects and languages the work is about.Question answering · 441 works

Cite

Cite as Chathurangi Shyalika, Dhaval C Patel and Amit Sheth. 2026. IndustryAssetEQA: A Neurosymbolic Operational Intelligence System for Embodied Question Answering in Industrial Asset Maintenance. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-industry.49

@inproceedings{shyalika2026industryasseteqa,
  title = {{IndustryAssetEQA: A Neurosymbolic Operational Intelligence System for Embodied Question Answering in Industrial Asset Maintenance}},
  author = {Chathurangi Shyalika, Dhaval C Patel and Amit Sheth},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Track)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-industry.49},
  url = {https://doi.org/10.18653/v1/2026.acl-industry.49}
}

References

41 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Arnab Biswas. 2020. Microsoft azure predictive maintenance. https://www.kaggle.com/datasets/ar nabbiswas1/microsoft-azure-predictive-mai ntenance. Accessed: 2026-02-01. ↗
  2. Joyjit Chatterjee and Nina Dethlefs. 2022. Automated question-answering for interactive decision support in operations & maintenance of wind turbines. IEEE Access, 10:84710–84737. find
  3. Jialin Chen, Aosong Feng, Ziyu Zhao, Juan Garza, Gaukhar Nurbek, Cheng Qin, Ali Maatouk, Leandros Tassiulas, Yifeng Gao, and Rex Ying. 2025. Mtbench: A multimodal time series benchmark for temporal reasoning and question answering. ArXiv, abs/2503.16858. — [arXiv:2503.16858](https://arxiv.org/abs/2503.16858) find
  4. Sihao Chen, Fan Zhang, Kazoo Sone, and Dan Roth. 2021. Improving faithfulness in abstractive summarization with contrast candidate generation and selection. In North American Chapter of the Association for Computational Linguistics.
  5. Christodoulos Constantinides, Shuxin Lin, and Dhaval C Patel. 2025a. Generalized embedding models for industry 4.0 applications. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing: Industry Track, pages 2234–2251. find
  6. Christodoulos Constantinides, Dhaval C Patel, Shuxin Lin, Claudio Guerrero, SUNIL DAGAJIRAO PATIL, and Jayant Kalagnanam. 2025b. Failuresensoriq: A multi-choice qa dataset for understanding sensor relationships and failure modes. In The Thirty-ninth Annual Conference on Neural Information Processing… find
  7. Pascale Fung, Yoram Bachrach, Asli Celikyilmaz, Kamalika Chaudhuri, Delong Chen, Willy Chung, Emmanuel Dupoux, Hongyu Gong, Hervé Jégou, Alessandro Lazaric, Arjun Majumdar, Andrea Madotto, Franziska Meier, Florian Metze, Louis-Philippe Morency, Théo Moutakanni, Juan Pino, Basile Terver, Joseph Tighe… — [arXiv:2506.22355](https://arxiv.org/abs/2506.22355) find
  8. Muhammad Fadhil Ginting, Dong-Ki Kim, Xiangyun Meng, Andrzej Marek Reinke, Bandi Jai Krishna, Navid Kayhani, Oriana Peltzer, David Fan, Amirreza Shaban, Sung-Kyun Kim, and 1 others. Enter the mind palace: Reasoning and planning for long-term active embodied question answering. In 9th Annual Conferen… find
  9. Nikolai Helwig, Eliseo Pignanelli, and Andreas Schütze. 2015a. Condition monitoring of a complex hydraulic system using multivariate statistics. In 2015 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) Proceedings, pages 210– 215. IEEE. find
  10. Nikolai Helwig, Eliseo Pignanelli, and Andreas Schütze. 2015b. Condition monitoring of hydraulic systems. UCI Machine Learning Repository. Accessed: 2026-01-30. find
  11. Nikolai Helwig, Eliseo Pignanelli, and Andreas Schütze. 2015c. D8. 1-detecting and compensating sensor faults in a hydraulic condition monitoring system. Proceedings SENSOR 2015, pages 641–646. find
  12. Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. 2023. Survey of hallucination in natural language generation. ACM computing surveys, 55(12):1–38.
  13. Yaxuan Kong, Yiyuan Yang, Yoontae Hwang, Wenjie Du, Stefan Zohren, Zhangyang Wang, Ming Jin, and Qingsong Wen. 2025. Time-mqa: Time series multi-task question answering with context enhancement. In Annual Meeting of the Association for Computational Linguistics.
  14. Laerd Statistics. 2018. Mcnemar’s test using spss statistics. https://statistics.laerd.com/spss-tut orials/mcnemars-test-using-spss-statistic s.php. Accessed: 2026-04-20. ↗
  15. Jay Lee, Jaskaran Singh, Moslem Azamfar, and Vibhor Pandhare. 2020. Industrial ai and predictive analytics for smart manufacturing systems. In Smart manufacturing, pages 213–244. Elsevier. find
  16. Sang-Hyuk Lee, So Won Choi, and Eul-Bum Lee. 2023. A question-answering model based on knowledge graphs for the general provisions of equipment purchase orders for steel plants maintenance. Electronics. find
  17. Yifan Li, Yuhang Chen, Anh Dao, Lichi Li, Zhongyi Cai, Zhen Tan, Tianlong Chen, and Yu Kong. 2025a. Industryeqa: Pushing the frontiers of embodied question answering in industrial scenarios. In Proceedings of the Thirty-ninth Annual Conference on Neural Information Processing Systems Datasets and Be… find
  18. Yifan Li, Lichi Li, Anh Dao, Xinyu Zhou, Yicheng Qiao, Zheda Mai, Daeun Lee, Zichen Chen, Zhen Tan, Mohit Bansal, and Yu Kong. 2025b. Industrynav: Exploring spatial reasoning of embodied agents in dynamic industrial navigation. ArXiv, abs/2511.17384. — [arXiv:2511.17384](https://arxiv.org/abs/2511.17384) find
  19. Zechuan Li, Hongshan Yu, Yihao Ding, Yan Li, Yong He, and Naveed Akhtar. 2026. Embodied intelligence for 3d understanding: A survey on 3d scene question answering. Information Fusion, 126:103624. find
  20. Shuxin Lin, Dhaval Patel, and Christodoulos Constantinides. 2025. Fine-tuned thoughts: Leveraging chain-of-thought reasoning for industrial asset health monitoring. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 20687– 20700. find
  21. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.
  22. Arjun Majumdar, Anurag Ajay, Xiaohan Zhang, Pranav Putta, Sriram Yenamandra, Mikael Henaff, Sneha Silwal, Paul Mcvay, Oleksandr Maksymets, Sergio Arnaud, Karmesh Yadav, Qiyang Li, Ben Newman, Mohit Sharma, Vincent-Pierre Berges, Shiqi Zhang, Pulkit Agrawal, Yonatan Bisk, Dhruv Batra, and 5 others. 2… find
  23. MCP Analytics. 2025. Mcnemar’s test: Practical guide for data-driven decisions. https://mcpanalyti cs.ai/articles/mcnemars-test-practical-g uide-for-data-driven-decisions. Published: December 26, 2025; Accessed: 2026-04-20. find
  24. Mike A Merrill, Mingtian Tan, Vinayak Gupta, Thomas Hartvigsen, and Tim Althoff. 2024. Language models still struggle to zero-shot reason about time series. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 3512–3533. find
  25. NASA Open Data Portal. 2025. C-MAPSS Jet Engine Simulated Data. https://data.nasa.gov/da taset/cmapss-jet-engine-simulated-data. Accessed: 2026-02-01. ↗
  26. Dhaval Patel, Shuxin Lin, James Rayfield, Nian-jun Zhou, Chathurangi Shyalika Jayakody, Suryanarayana R Yarrabothula, Roman Vaculin, Natalia Martinez, Fearghal O’Donncha, and Jayant Kalagnanam. 2025. Assetopsbench: A real-world evaluation benchmark for ai-driven task automation in industrial asset m… — [arXiv:2506.03828](https://arxiv.org/abs/2506.03828) find
  27. Ricardo Silva Peres, Xiaodong Jia, Jay Lee, Keyi Sun, Armando Walter Colombo, and Jose Barata. 2020. Industrial artificial intelligence in industry 4.0-systematic review, challenges and outlook. IEEE access, 8:220121–220139. find
  28. RDFLib Developers. 2026. Rdflib: A python library for working with rdf. https://github.com/RDFLib/ rdflib. Version 7.6.0, Accessed: 2026-04-20. ↗
  29. Abhinav Saxena, Kai Goebel, Don Simon, and Neil Eklund. 2008. Damage propagation modeling for aircraft engine run-to-failure simulation. In 2008 international conference on prognostics and health management, pages 1–9. IEEE. find
  30. Patrícia Schmidtová, Eduardo Calò, Simone Balloccu, Dimitra Gkatzia, Rudali Huidrom, Mateusz Lango, Fahime Same, Vilém Zouhar, Saad Mahamood, and Ondřej Dušek. 2025. Do my eyes deceive me? a survey of human evaluations of hallucinations in nlg. In Proceedings of the 18th International Natural Langu… find
  31. Chaoran Zhang, Chenhao Zhang, Zhaobo Xu, Qinghongbing Xie, Pingfa Feng, and Long Zeng. 2025. Embodied intelligent industrial robotics: Concepts and techniques. ArXiv, abs/2505.09305. — [arXiv:2505.09305](https://arxiv.org/abs/2505.09305) find
  32. Tizian Schneider, Steffen Klein, and Manuel Bastuck. 2018. Condition Monitoring of Hydraulic Systems Data Set. Accessed: 2026-01-30. find
  33. Dammika Seneviratne, Lorenzo Ciani, Marcantonio Catelani, Diego Galar, and 1 others. 2018. Smart maintenance and inspection of linear assets: An industry 4.0 approach. Acta Imeko, 7:50–56. find
  34. Amit Sheth, Kaushik Roy, and Manas Gaur. 2023. Neurosymbolic artificial intelligence (why, what, and how). IEEE Intelligent Systems, 38(3):56–62. find
  35. Chen Su, Yuanhe Tian, and Yan Song. 2025. Chain-of-thought reviewing and correction for time series question answering. ArXiv, abs/2512.22627. find
  36. Sinan Tan, Mengmeng Ge, Di Guo, Huaping Liu, and Fuchun Sun. 2021. Knowledge-based embodied question answering. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45:11948–11960. find
  37. Vagan Y. Terziyan, Oleksandra Vitko, and Oleksandr Terziyan. 2025. A conceptual design of industrial asset maintenance system by autonomous agents enhanced with chatgpt. Intelligent and Sustainable Manufacturing. find
  38. Md Nayem Uddin, Amir Saeidi, Divij Handa, Agastya Seth, Tran Cao Son, Eduardo Blanco, Steven Corman, and Chitta Baral. 2024. Unseentimeqa: Time-sensitive question-answering beyond llms’ memorization. In Annual Meeting of the Association for Computational Linguistics.
  39. Alexander von Birgelen and Oliver Niggemann. 2018. Anomaly detection and localization for cyber-physical production systems with self-organizing maps. In IMPROVE-Innovative Modelling Approaches for Production Systems to Raise Validatable Efficiency: Intelligent Methods for the Factory of the Future,… find
  40. Yilin Wang, Peixuan Lei, Jie Song, Yuzhe Hao, Tao Chen, Yuxuan Zhang, Lei Jia, Yuanxiang Li, and Zhongyu Wei. 2025. Itformer: Bridging time series and natural language for multi-modal qa with large-scale multitask dataset. In 42nd International Conference on Machine Learning, volume abs/2506.20093. — [arXiv:2506.20093](https://arxiv.org/abs/2506.20093) find
  41. Hong Yung Yip and Amit P. Sheth. 2024. The empwr platform: Data and knowledge-driven processes for the knowledge graph lifecycle. IEEE Internet Computing, 28:61–69. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/shyalika-2026-industryasseteqa-neurosymbolic-operational-intelligence-system/