KYHTQT.

Enhancing the Performance of Vietnamese Online Public Service Chatbots with RAG

Năm XB 2024 Tạp chí / Hội thảo Lecture Notes in Networks and Systems Volume 1205 LNNS Đơn vị CNTT DOI / Link https://doi.org/10.1007/978-3-031-80943-9_8 ↗

Tác giả

Trung-Nghia Phung

Tóm tắt

Public service chatbots play a crucial role in providing efficient and accessible services to citizens, but their performance often falls short in specialized domains such as public administration. This paper presents an approach to enhance the performance of Vietnamese public service chatbots by integrating Retrieval-Augmented Generation (RAG). By leveraging Langchain’s flexible architecture, the chatbot can seamlessly make API calls and interact with databases, ensuring access to the most relevant and current information. Especially, we explored RAG with large language models like Vistral-7B-chat, combined with retrieval mechanisms, to provide contextually appropriate responses. Our experimental results demonstrate a significant improvement in response accuracy, addressing the limitations of generic open-domain models. This enhancement offers substantial benefits for public administrative interactions, significantly improving efficiency and quality in governmental applications.

Tài liệu tham khảo

[1] Følstad, A., et al.: Future directions for chatbot research: an interdisciplinary research agenda. Computing 103 (2021). https://doi.org/10.1007/s00607-021-01016-7

[2] Jiang, A.Q., et al.: Mistral 7b (2023)

[3] Kulkarni, M., Tangarajan, P., Kim, K., Trivedi, A.: Reinforcement learning for optimizing rag for domain chatbots (2024)

[4] Kumar, V., et al.: Large-language-models (llm)-based ai chatbots: Architecture, in-depth analysis andÂ their performance evaluation. In: Santosh, K., et al. (eds.) Recent Trends in Image Processing and Pattern Recognition, pp. 237–249. Springer, Cham (2024). https://doi.org/10.1007/978-3-031-53085-2_20

[5] Lewis, P., et al.: Retrieval-augmented generation for knowledge-intensive nlp tasks. In: Proceedings of the 34th International Conference on Neural Information Processing Systems. NIPS ’20. Curran Associates Inc., Red Hook (2020)

[6] Nguyen, D.Q., Nguyen, A.T.: PhoBERT: pre-trained language models for Vietnamese. In: Findings of the Association for Computational Linguistics: EMNLP 2020, pp. 1037–1042 (2020)

[7] Setty, S., Jijo, K., Chung, E., Vidra, N.: Improving retrieval for rag based question answering models on financial documents (2024)

[8] Touvron, H., et al.: Llama: open and efficient foundation language models. ArXiv arxiv:2302.13971 (2023). https://api.semanticscholar.org/CorpusID:257219404

[9] Wang, X., Dai, H., Gao, S., Li, P.: Characteristic AI agents via large language models. In: Calzolari, N., Kan, M.Y., Hoste, V., Lenci, A., Sakti, S., Xue, N. (eds.) Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pp. 3016–3027. ELRA and ICCL, Torino (2024). https://aclanthology.org/2024.lrec-main.269

Ghi chú

ICTA

← Quay lại danh sách bài báo