Page 337 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

3 Use Case Requirements

• REQ-1: It is critical to fine-tune the telecom business acceptance large language model
with internal telecommunication carriers' data, including telecom product and service
information and business processes, which enable the large language model to master 4.3 - 5G
domain-specific knowledge of telecom service operations.
• REQ-2: It is of added value to continuously and iteratively optimise capabilities to adapt
to evolving telecom products and business processes.

4 Sequence Diagram

5 References

[1] H. Zou, Q. Zhao, Y. Tian, L. Bariah, F. Bader, T. Lestable, and M. Debbah, "TelecomGPT:
A Framework to Build Telecom-Specific Large Language Models," arXiv preprint
arXiv:2407.09424, Jul. 2024. [Online]. Available: https:// arxiv .org/ abs/ 2407 .09424
[2] G. M. Yilma, J. A. Ayala-Romero, A. Garcia-Saavedra, and X. Costa-Perez, "TelecomRAG:
Taming Telecom Standards with Retrieval Augmented Generation and LLMs," arXiv
preprint arXiv:2406.07053, Jun. 2024. [Online]. Available: https:// arxiv .org/ abs/ 2406
.07053
[3] T.-E. Lin, Y. Wu, F. Huang, L. Si, J. Sun, and Y. Li, "Duplex Conversation: Towards Human-
like Interaction in Spoken Dialogue Systems," arXiv preprint arXiv:2205.15060, May 2022.
[Online]. Available: https:// arxiv .org/ abs/ 2205 .15060
[4] Mixture of Experts Explained, Hugging Face Blog, Jun. 2023. [Online]. Available: https://
huggingface .co/ blog/ moe
[5] DeepSeek AI, “deepseek-ai/DeepSeek-R1-Distill-Qwen-32B,” GitHub, 2025. [Online].
Available: https:// huggingface .co/ deepseek -ai/ DeepSeek -R1 -Distill -Qwen -32B

301