Schema Aware Conversational Data Collection

• Lucas Emmanuel de Sousa Alves Universidade Federal de Campina Grande (UFCG)
• Claudio E. C. Campelo Universidade Federal de Campina Grande (UFCG)

DOI: https://doi.org/10.5753/sbbd.2026.249461

Resumo

This paper proposes a conversational approach to structured data collection using an automated agent that addresses limitations of traditional form-based methods. Guided by a predefined schema of attributes, types, and constraints, the agent conducts dialogues to clarify ambiguities, complete missing information, and validate data in real time while minimizing the number of interactions. To support this, we developed a Python library enabling rapid creation of applications that proactively request user input. Using this library, we built a chatbot evaluated through user-based experiments in a culinary habits case study. Results show the approach effectively validates constraints, supports semantic inference, and guides users to provide complete and consistent data.

Palavras-chave: Conversational Data Collection, Schema-Guided Dialogue, Multi-Agent Systems, Dialogue State Tracking, Data Validation, Structured Information Extraction, LLMs

Referências

Guest, G., Bunce, A., and Johnson, L. (2006). How many interviews are enough? Field Methods - FIELD METHOD, 18:59–82.

Henderson, M., Thomson, B., and Williams, J. D. (2014). The second dialog state tracking challenge. In Georgila, K., Stone, M., Hastie, H., and Nenkova, A., editors, Proceedings of the 15th Annual Meeting of the Special Interest Group on Discourse and Dialogue (SIGDIAL), pages 263–272, Philadelphia, PA, U.S.A. Association for Computational Linguistics.

Luo, Y., Ru, X., Liu, K., Yuan, L., Sun, M., Zhang, N., Liang, L., Zhang, Z., Zhou, J., Wei, L., Zheng, D., Wang, H., and Chen, H. (2025). Oneke: A dockerized schema-guided llm agent-based knowledge extraction system. In Companion Proceedings of the ACM on Web Conference 2025, WWW ’25, page 2871–2874, New York, NY, USA. Association for Computing Machinery.

Mehri, S. and Eskénazi, M. (2021). Schema-guided paradigm for zero-shot dialog. CoRR, abs/2106.07056.

Rastogi, A., Zang, X., Sunkara, S., Gupta, R., and Khaitan, P. (2020). Schema-guided dialogue state tracking task at DSTC8. CoRR, abs/2002.01359.

Tavares, D., Semedo, D., Rudnicky, A., and Magalhaes, J. (2023). Learning to ask questions for zero-shot dialogue state tracking. In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR ’23, page 2118–2122, New York, NY, USA. Association for Computing Machinery.

Wei, X., Cui, X., Cheng, N., Wang, X., Zhang, X., Huang, S., Xie, P., Xu, J., Chen, Y., Zhang, M., Jiang, Y., and Han, W. (2024). Chatie: Zero-shot information extraction via chatting with chatgpt.

Xu, D., Chen, W., Peng, W., Zhang, C., Xu, T., Zhao, X., Wu, X., Zheng, Y., Wang, Y., and Chen, E. (2024a). Large language models for generative information extraction: A survey.

Xu, L., Peng, N., Zhou, D., Ng, S.-K., and Fu, J. (2024b). Chain of thought explanation for dialogue state tracking.

Zhang, N., Xu, X., Tao, L., Yu, H., Ye, H., Qiao, S., Xie, X., Chen, X., Li, Z., and Li, L. (2022). DeepKE: A deep learning based knowledge extraction toolkit for knowledge base population. In Che, W. and Shutova, E., editors, Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 98–108, Abu Dhabi, UAE. Association for Computational Linguistics.

Zhao, J., Gupta, R., Cao, Y., Yu, D., Wang, M., Lee, H., Rastogi, A., Shafran, I., and Wu, Y. (2022). Description-driven task-oriented dialog modeling. CoRR, abs/2201.08904.