This research introduces a groundbreaking semantic communication system that leverages Multi-modal Large Language Models to enable more intelligent and efficient data transmission across networks.

• Develops a specialized MLLM architecture optimized for semantic communication
• Implements a novel system supporting multiple modalities, tasks, and users simultaneously
• Achieves significant improvements in transmission efficiency through semantic extraction
• Incorporates innovative security features through semantic separation mechanisms

This engineering advancement represents a fundamental shift in how we approach network communications, moving beyond traditional bit-based transmission to semantic-centered approaches that could dramatically improve bandwidth utilization while enabling richer multi-modal interactions.

M4SC: An MLLM-based Multi-modal, Multi-task and Multi-user Semantic Communication System