AI/LLMs Based Intelligent Telecommunications

Motivation

As telecommunications systems evolve toward AI-native and autonomous networks, there is a growing need for intelligent systems capable of understanding, reasoning, and acting across complex operational environments. Generative AI and large language models provide a powerful foundation for this transformation by combining language understanding, domain knowledge, and reasoning capabilities within a single adaptable framework.

Project Goals

• Leverage large language models for better communication system operation.
• Develop AI/LLM-native next-generation communication systems.

Current Technical Direction

LLM-Based Communication Traffic Forecasting

We aim to leverage multiple LLM agents to dynamically adjust model parameters and forecasting strategies, improving forecasting accuracy and robustness under changing network conditions.

LLM-Based Network Slicing

In intent-driven networks, operators often express service requirements using high-level objectives rather than explicit optimization formulations. LLMs can translate these intents into slice-management actions, facilitating autonomous and context-aware network slicing for diverse applications and users.

LLM-Based Communication Digital Twin

We aim to develop an LLM-based communication digital twin framework that provides a semantic and knowledge-driven way to understand, forecast, and optimize communication network behavior.