ZTE Communications Special Issue on
Multimodal Semantic Communication Technologies
The rapid evolution of 6G and beyond is shifting the paradigm of communication systems from bit-level transmission to semantic-level understanding. With the proliferation of multimodal applications involving text, images, audio, video, and sensory data, future communication networks should efficiently transmit, align, and interpret heterogeneous information across multiple modalities. Unlike conventional communication systems that primarily focus on accurate bit reconstruction, multimodal semantic communication aims to extract and transmit task-relevant semantic information while preserving cross-modal consistency and complementary relationships, thereby improving communication efficiency, reliability, and adaptability.
Recent advances in Large Language Models (LLMs), Large Vision Models (LVMs), and Agentic AI have opened new opportunities for multimodal semantic representation, compression, transmission, and reconstruction. By exploiting semantic correlations and complementary information across modalities, intelligent communication systems can enable more efficient information exchange under limited bandwidth and dynamic network conditions. However, realizing multimodal semantic communication still faces fundamental challenges in joint source-channel coding (JSCC), cross-modal alignment, semantic fidelity evaluation, and robust semantic reconstruction. Furthermore, supporting distributed multimodal applications requires efficient edge-cloud collaboration and adaptive semantic transmission mechanisms.
This special issue invites original research on novel architectures, algorithms, and experimental validations that advance Multimodal Semantic Communication. We particularly welcome submissions addressing fundamental challenges in cross-modal semantic representation, efficient semantic transmission, and reliable semantic reconstruction for next-generation wireless networks.
This special issue seeks original articles on Multimodal Semantic Communication Technologies. Topics of interest include, but are not limited to:
* Multimodal representation learning and semantic encoding for communication
* Cross-modal semantic alignment, translation, and fusion techniques
* Semantic-aware channel coding and modulation for multimodal data
* Token communication and efficient token-level semantic transmission
* Task-oriented multimodal semantic communication toward edge intelligence and IoT
* Generative AI and large foundation models for multimodal semantic compression and reconstruction
* Agentic AI and multi-agent collaboration for multimodal semantic communication
* Retrieval-augmented and knowledge-enhanced multimodal semantic communication
* Knowledge base construction and evolution for multimodal semantic communication
* Semantic-aware resource allocation and network optimization for multimodal communications
* Edge deployment and performance optimization of lightweight multimodal semantic models
* Multimodal semantic communication for collaborative embodied intelligence
* Human-centric semantic communication for immersive and XR applications
* Privacy, security, and adversarial robustness in multimodal semantic communication
* Multimodal semantic transmission in space-air-ground-sea integrated networks
This special issue aims to gather cutting-edge contributions in the field of multimodal semantic communication, and promote the deep integration of agentic AI, multimodal learning and next-generation communication networks. We warmly invite researchers and engineers worldwide to submit high-quality papers, jointly explore the theoretical boundaries and technical pathways of semantic communication, and provide new ideas, methods and support for the development of 6G and future intelligent communication networks. All submissions will undergo rigorous peer review and accepted papers will be published in the special issue.
Zhang Ping, Beijing University of Posts and Telecommunications, China
Ma Nan, Beijing University of Posts and Telecommunications, China (manan@bupt.edu.cn)
l
Manuscripts must be written in English and submitted electronically in MS Word or a compatible format. The recommended length is between 3 000 and 8 000 words, with no more than 10 figures and tables included. Authors should provide mathematical content and graphics in editable formats.
Please submit your paper via the journal’s online submission portal at https://mc03.manuscriptcentral.com/ztecom.
