Scopus

Beyond Monolithic LLMs: A Hybrid Framework for Robust Natural Language to Data Visualization Generation

Năm XB 2026 Tạp chí / Hội thảo International Conference on Advances in Information and Communication Technology (ICTA2025) Vol 2, 327–335 Đơn vị CNTT DOI / Link https://doi.org/10.1007/978-3-032-18159-6_35 ↗

Tác giả

Thi Minh-Hue Luong ; The-Vinh Nguyen ✉ ; Duc-Quang Vu ; Van-Viet Nguyen ; Huu-Khanh Nguyen ; Kim-Son Nguyen ; Xuan-Truong Quach

Tóm tắt

Data visualization generation from natural language queries (NL2VIS) has emerged as a critical research area, enabling non-technical users to create insights from structured data. While recent approaches have leveraged Large Language Models (LLMs) for end-to-end NL2VIS tasks, they suffer from the “lost in the middle” problem, generating incorrect visualization grammar for complex queries and requiring extensive post processing of unstructured outputs. This study introduces a hybrid agent framework that decomposes NL2VIS into structured subtasks, combining LLM understanding with formal grammar constraints and schema validation. Experimental evaluation on the VisEval benchmark dataset demonstrates significant performance improvements, with our agent-based system achieving 50.64% exact match accuracy and 99.74% execution success rate, surpassing existing LLM-based approaches by 18.6 …

Tài liệu tham khảo

[1] Chen, N., Zhang, Y., Xu, J., Ren, K., Yang, Y.: VisEval: a benchmark for data visualization in the era of large language models. IEEE Trans. Vis. Comput. Graph. (2024)

[2] Dibia, V.: LIDA: a tool for automatic generation of grammar-agnostic visualizations and infographics using large language models. In: The 61st Annual Meeting of The Association for Computational Linguistics (2023)

[3] Gu, J., et al.: A survey on LLM-as-a-judge. arXiv preprint arXiv:2411.15594 (2024)

[4] Hue, L.T.M., Vinh, N.T.: Overview of the application of generative artificial intelligence in data visualization problems. Vinh Univ. J. Sci. (2025)

[5] Luo, Y., Tang, N., Li, G., Chai, C., Li, W., Qin, X.: Synthesizing natural language to visualization (NL2VIS) benchmarks from NL2SQL benchmarks. In: Proceedings of the 2021 International Conference on Management of Data, pp. 1235–1247 (2021)

[6] Luo, Y., Tang, N., Li, G., Tang, J., Chai, C., Qin, X.: Natural language to visualization by neural machine translation. IEEE Trans. Visual Comput. Graphics 28(1), 217–226 (2021)

[7] Luong-Thi-Minh, H., Nguyen-The, V., Xuan, T.Q.: VizAgent: towards an intelligent and versatile data visualization framework powered by large language models. In: International Conference on Advances in Information and Communication Technology, pp. 89–97. Springer (2024)

[8] Maddigan, P., Susnjak, T.: Chat2VIS: generating data visualizations via natural language using chatGPT, codex and gpt-3 large language models. IEEE Access 11, 45181–45193 (2023)

[9] Ouyang, G., et al.: nvAgent: automated data visualization from natural language via collaborative agent workflow. arXiv preprint arXiv:2502.05036 (2025)

[10] Shen, L., et al.: Towards natural language interfaces for data visualization: a survey. IEEE Trans. Visual Comput. Graphics 29(6), 3121–3144 (2022)

[11] Tian, Y., et al.: ChartGPT: leveraging LLMs to generate charts from abstract natural language. IEEE Trans. Vis. Comput. Graph. (2024)

[12] Wang, Q., Chen, Z., Wang, Y., Qu, H.: A survey on ML4VIS: applying machine learning advances to data visualization. IEEE Trans. Visual Comput. Graphics 28(12), 5134–5153 (2021)

[13] Wu, Y., et al.: Automated data visualization from natural language via large language models: an exploratory study. Proc. ACM Manag. Data 2(3), 1–28 (2024)

[14] Zhang, J., Zhang, R., Kong, F., Miao, Z., Ye, Y., Zheng, Y.: Lost-in-the-middle in long-text generation: synthetic dataset, evaluation framework, and mitigation. arXiv preprint arXiv:2503.06868 (2025)

← Quay lại danh sách bài báo