Literature Survey: Semantic Communication

Domain: Future Networks & 6G
Topic Search: semantic communication
Timeframe: 2021 - 2026

This is a curated survey of recent publications focusing on semantic communication. Results are filtered for top-tier journals and prominent conferences.

📚 Curated Peer-Reviewed Publications

1. Perception-Enhanced Multitask Multimodal Semantic Communication for UAV-Assisted Integrated Sensing and Communication System

Venue: ICC | Year: 2025 | Citations: 4 Authors: Ziji Guo, Haonan Tong, Zhilong Zhang, Danpu Liu

Recent advances in integrated sensing and communication (ISAC) unmanned aerial vehicles (UAVs) have enabled their widespread deployment in critical applications such as emergency management. This paper investigates the challenge of efficient multitask multimodal data communication in UAV-assisted ISAC systems, in the considered system model, hyperspectral (HSI) and LiDAR data are collected by UAV-mounted sensors for both target classification and data reconstruction at the terrestrial BS. The limited channel capacity and complex environmental conditions pose significant challenges to effective air-to-ground communication. To tackle this issue, we propose a perception-enhanced multitask multimodal semantic communication (PE-MMSC) system that strategically leverages the onboard computational and sensing capabilities of UAVs. In particular, we first propose a robust multimodal feature fusion method that adaptively combines HSI and LiDAR semantics while considering channel noise and task requirements. Then the method introduces a perception-enhanced (PE) module incorporating attention mechanisms to perform coarse classification on UAV side, thereby optimizing the attention-based multimodal fusion and transmission. Experimental results demonstrate that the proposed PE-MMSC system achieves 5%–10% higher target classification accuracy compared to conventional systems without PE module, while maintaining comparable data reconstruction quality with acceptable computational overheads.

2. A Goal-Oriented Context-Aware Adaptive Semantic Communication Scheme Using a Semantic Mask Module

Venue: ICC | Year: 2025 | Citations: 1 Authors: Oladayo Ogunjimi, Fangzhe Chen, Fang Fang 0005, Xianbin Wang 0001

By extracting and transmitting semantic information (SI), semantic communications (SC) can achieve communication goal and reliable transmission with much lower data rate among devices. However, devices in SC typically need to perform concurrent tasks, subject to varying available computational resources. Additionally, dynamic communication environments could dramatically deteriorate communication performance. To achieve robust SC with aforementioned uncertainties, this paper proposes a context-aware, adaptive rate, multitask SC scheme (CAAR-MTSC) to optimize the latent representation of transmitted data. The proposed scheme comprehensively considers the user’s perceptual needs, current channel conditions, and task-relevant information within the data. Specifically, we introduce a Semantic Mask Module (SMM) that dynamically controls the data rate based on channel conditions and user-defined perceptual metrics to obtain the best trade-off between task performance and data rate. To enhance multi-task performance, we formulate a triple trade-off information bottleneck (IB) optimization problem using rate-distortion perception. We further incorporate a knowledge distillation (KD) strategy to reduce the model size while maintaining performance. Simulation results indicate that, in low-SNR conditions, the proposed framework achieves up to a 12% improvement in SSIM and a 3.4% increase in classification accuracy while reducing data rate by 11%, thereby demonstrating its efficacy under resource-constrained scenarios.

3. Semantic Knowledge Base Based Dual-mode Video Semantic Communication

Venue: ICC | Year: 2026 | Citations: 0 Authors: Zhicheng Bao, Long Liu, Nan Ma 0014, Chen Dong 0001, Hao Chen 0013 et al.

Semantic communication is a highly promising field designed to boost efficiency by prioritizing the transfer of meaning over raw bit fidelity. Existing research generally focuses on a single transmission mechanism, operating exclusively within either the continuous or discrete semantic space. However, a significant gap remains: there is no unified framework that successfully integrates the benefits of both modes. To overcome this limitation, and specifically targeting video data, this paper introduces a dual-mode video semantic communication (DMVSC) framework that seamlessly combines the continuous and discrete approaches. To further improve transmission efficiency, we design a temporal redundancy compressor based on an attention mechanism. Additionally, we construct a specialized semantic knowledge base (SKB) for video features, enabling efficient semantic feature quantization via index transmission. Experimental results conclusively demonstrate that our proposed DMVSC framework achieves superior video reconstruction performance compared to conventional methods.

4. Robust Multimodal Semantic Communications with Semantic Fusion and Compensation

Venue: ICC | Year: 2026 | Citations: 0 Authors: Xiang Peng, Zhijin Qin, Xiaoming Tao 0001, Jianhua Lu

Semantic communication has demonstrated significant potential to improve communication network performance. However, ensuring robustness in multimodal semantic communication remains challenging due to both physical and semantic impairments, with the latter largely underexplored. In this paper, we investigate multimodal semantic impairments and propose a novel framework for robust multimodal semantic communications. We introduce a performance metric to quantify semantic impairments and construct a corresponding dataset for validation. Moreover, we design DeepSC-RM, a deep learning enabled semantic communication system, which integrates an adaptive multimodal semantic fusion module to dynamically combine information across modalities, a multimodal joint semantic–channel codec for symbol-level transmission, and a cross-modal semantic compensation module at the receiver to enhance accurate semantic interpretation. Experimental results demonstrate the superior performance of the proposed DeepSC-RM, confirming its effectiveness in mitigating semantic impairments and achieving robust multimodal transmission.

5. Multitask Semantic-Coded Image Communication for UAV Integrated Sensing and Communication: A Channel-wise Feature Enhancement Approach

Venue: INFOCOM | Year: 2026 | Citations: 0 Authors: Chen Mao, Zhongqiang Zhang, Shuhang Zhang, Shuai Ma 0002, Guangming Shi et al.

In integrated sensing and communication (ISAC) unmanned aerial vehicle (UAV) networks, the single-to-noise ratio (SNR) and available data rate continuously vary with altitude, distance, blockage, and maneuvering, which heavily challenges conventional image compression and channel coding methods with layered and separate design. In this paper, we propose a multitask semantic-coded image communication (MSCIC) framework for UAV-ISAC with a channel-wise feature enhancement (CFE) mechanism, which could leverage involution and adjustment blocks to emphasize robust informative features under varying channel states. Besides, a transmission-rate adaptive design is developed to dynamically control the amount of transmitted features through a rate adjustment module, together with a corresponding decoding module to support multitask inference. The simulation results show that the proposed framework performs better with adaptive UAV image transmission in varying SNRs compared with typical algorithms.

6. Semantic Knowledge Base-Enhanced Joint Source-Channel Coding Framework for Robust Semantic Communications

Venue: WCNC | Year: 2026 | Citations: 0 Authors: Haixiao Gao, Mengying Sun, Yanhan Wang, Xiaodong Xu 0001, Zechuan Fang et al.

In this paper, we propose a semantic knowledge base (SKB)-enhanced joint source-channel coding (JSCC) framework, named SKB-JSCC. This framework splits semantic features into two transmission branches with dynamically configurable split ratio, a JSCC-based branch that provides superior performance at high signal-to-noise ratio (SNR), and an SKB-based branch that strengthens robustness at low SNR. Additionally, we introduce a bidirectional gated mutual errorcorrection (BGMEC) module at the receiver and provide theoretical analysis supporting its soundness and effectiveness. The BGMEC module uses cross attention to learn semantic associations between the two branches of streams and treats one stream as conditional information for the other to realize mutual correction and feature fusion. Experimental results illustrate that the BGMEC module provides performance gains for SKB-JSCC over the entire SNR range. Compared with a JSCC-based baseline, SKB-JSCC achieves a 14.3% improvement in transmission performance at low SNR and effectively mitigates the “cliff effect” relative to traditional separation-based schemes.

7. TRILLM: Triple-Based Semantic Communication with LLMS for Robust IIoT Communication

Venue: WCNC | Year: 2026 | Citations: 0 Authors: Justin Jose, Shanthakumar Karan, A. S. Madhukumar

This paper presents a triple-based large language model (LLM)-to-LLM (TRILLM) semantic communication framework for handling semantically impaired industrial internet of things (IIoT) text. We consider a system with $N$ edge devices where each device uses a lightweight LLM to compress event-driven messages into triple representations before transmission to a central hub equipped with an LLMbased semantic decoder. To better evaluate reconstruction quality under semantic impairments, we introduce a semantic equivalence metric that combines cosine similarity with grammar fluency. We further analyze the rate and latency characteristics of the proposed framework. Simulation results show that TRILLM achieves up to 67% higher semantic equivalence than baseline knowledge graph (KG)-based schemes, demonstrating improved robustness in impaired IIoT text communication. Additionally, TRILLM offers adaptability and contextual reasoning advantages over KG approaches.

8. Semantic Communication-Enabled Cooperative BEV Semantic Segmentation Frame in Internet of Vehicles

Venue: WCNC | Year: 2026 | Citations: 0 Authors: Ning Xu, Puning Zhang, Lekang Ye, Guangqian Wang

The precision of Bird’s Eye View (BEV) semantic segmentation directly affects the decision-making reliability of autonomous driving systems. However, existing methods encounter significant challenges in feature redundancy suppression, channel robustness enhancement, and multi-vehicle feature alignment. To tackle these issues, this paper presents a Semantic Communication-Enabled Cooperative BEV Semantic Segmentation Framework (SCC-BEVSS). First, we establish a variational optimization objective grounded in Information Bottleneck (IB) theory to extract compact semantic features from multi-view images, effectively reducing feature redundancy in BEV representations. Second, a Joint Source-Channel Coding (JSCC) scheme is designed to ensure reliable transmission of semantic features under IoV low signal-to-noise ratio (SNR) conditions. Finally, we introduce a multi-level feature fusion mechanism that incorporates a feature space calibration module and an attention enhancement module to address spatial misalignment issues in multi-vehicle cooperation. Experimental results demonstrate that the proposed framework surpasses existing methods with notable gains in mean Intersection over Union (mIoU).

9. SemTexIB: Semantic Text Communication with Information Bottleneck: Integrating Rate and Semantic Similarity into Training Objectives

Venue: GLOBECOM | Year: 2025 | Citations: 0 Authors: Abdulrahman Alamoudi, Ahmet Faruk Saz, Yashas Malur Saidutta, Faramarz Fekri

Recent major developments in semantic communication systems stem from integration of deep learning (DL) techniques. Following the discovery of capacity achieving codes, the primary motivation for adopting the semantic approach, which retrieves meaning without requiring an exact reconstruction, is its potential to further conserve resources such as bandwidth and power. In this paper, we propose a novel semantic communication framework for textual data over additive white Gaussian noise (AWGN) channels via DL. Our framework leverages the information bottleneck (IB) principle to balance minimizing bit transmission under wireless channel rate constraints with maximizing semantic information retention. Unlike previous works, we integrate the bilingual evaluation understudy (BLEU) sentence similarity score into the training objective to enhance model performance. In particular, inspired by knowledge distillation, we utilize large language models (LLMs) during training to transfer their knowledge of text semantics into our model. Using IB principle, we train a neural semantic encoder at the transmitter and a neural semantic decoder at the receiver that incorporates into its objective function the rate constraint together with the BLEU score and the knowledge encoded in the soft probabilities produced by the LLM. Through extensive experiments, our proposed framework demonstrates a notable improvement of up to 45% in text semantic similarity compared to state-of-the-art benchmarks operating at the same channel capacity, significantly outperforming traditional communication systems. Moreover, it exhibits robustness to variations in signal-to-noise ratio (SNR) and achieves significant gains across both low and medium SNR regimes.

10. Digital Semantic Communication in ISAC: A Framework for Enhanced Sensing and Communication

Venue: GLOBECOM | Year: 2025 | Citations: 0 Authors: Ouwen Huan, Chuanhong Liu, Nuocheng Yang, Zhilong Zhang, Tao Luo 0005 et al.

This paper proposes a novel integrated sensing and communication (ISAC) framework incorporating digital semantic communication (SemCom) to resolve the tradeoff between sensing and communication performance. In particular, to accomplish task-oriented semCom, the base station (BS) extracts semantic symbols and transmits each dimension over different orthogonal frequency division multiplex (OFDM) subcarriers. To achieve sensing objective, the BS broadcasts OFDM signals and receives echoes via a uniform linear array (ULA) to estimate echo channel state information (CSI) and obtain target parameters. Given the varying task-related importance of each dimension, the framework allocates quantization bits, modulation order, and transmission power accordingly to meet SemCom requirements. On the other hand, sensing performance is evaluated using the Cramér-Rao Bound (CRB) of echo CSI, with transmission power allocation optimized to enhance sensing. The problem is formulated to minimize the sensing CRB while satisfying SemCom task loss, total resources, and transmission efficiency constraints. To solve this problem, we introduce a Hybrid Action Space Proximal Policy Optimization (H-PPO) algorithm, which can simultaneously determine the power allocated for each dimension from a continuous action space, and select a proper number of quantization bits and modulation order from discrete action spaces. Simulations show that the proposed method enhances SemCom task performance by up to 77% and reduces sensing error by up to 58% compared to conventional digital systems.

11. Collaborative Knowledge Sharing-Empowered Effective Semantic Rate Maximization for Two-Tier Semantic-Bit Communication Networks

Venue: ICC | Year: 2025 | Citations: 0 Authors: Hong Chen 0016, Fang Fang 0005, Xianbin Wang 0001

Effective task-oriented semantic communications relies on perfect knowledge alignment between transmitters and receivers for accurate recovery of task-related semantic information, which can be susceptible to knowledge misalignment and performance degradation in practice. To tackle this issue, continual knowledge updating and sharing are crucial to adapt to evolving task and user related demands, despite the incurred resource overhead and increased latency. In this paper, we propose a novel collaborative knowledge sharing-empowered semantic transmission mechanism in a two-tier edge network, exploiting edge cooperations and bit communications to address KB mismatch. By deriving a generalized effective semantic transmission rate (GESTR) that considers both semantic accuracy and overhead, we formulate a mixed integer nonlinear programming problem to maximize GESTR of all mobile devices by optimizing knowledge sharing decisions, extraction ratios, and BS/subchannel allocations, subject to task accuracy and delay requirements. The joint optimum solution can be obtained by proposed fractional programming based branch and bound algorithm and modified Kuhn-Munkres algorithm efficiently. Simulation results demonstrate the superior performance of proposed solution, especially in low signal-to-noise conditions.

⚡ Latest Pre-Prints

1. Semantic Revolution from Communications to Orchestration for 6G: Challenges, Enablers, and Research Directions

Published: 2024-06-24 Authors: Masoud Shokrnezhad, Hamidreza Mazandarani, Tarik Taleb, Jaeseung Song, Richard Li

In the context of emerging 6G services, the realization of everything-to-everything interactions involving a myriad of physical and digital entities presents a crucial challenge. This challenge is exacerbated by resource scarcity in communication infrastructures, necessitating innovative solutions for effective service implementation. Exploring the potential of Semantic Communications (SemCom) to enhance point-to-point physical layer efficiency shows great promise in addressing this challenge. However, achieving efficient SemCom requires overcoming the significant hurdle of knowledge sharing between semantic decoders and encoders, particularly in the dynamic and non-stationary environment with stringent end-to-end quality requirements. To bridge this gap in existing literature, this paper introduces the Knowledge Base Management And Orchestration (KB-MANO) framework. Rooted in the concepts of Computing-Network Convergence (CNC) and lifelong learning, KB-MANO is crafted for the allocation of network and computing resources dedicated to updating and redistributing KBs across the system. The primary objective is to minimize the impact of knowledge management activities on actual service provisioning. A proof-of-concept is proposed to showcase the integration of KB-MANO with resource allocation in radio access networks. Finally, the paper offers insights into future research directions, emphasizing the transformative potential of semantic-oriented communication systems in the realm of 6G technology.

2. A Survey on Semantic Communications in Internet of Vehicles

Published: 2025-03-03 Authors: Sha Ye, Qiong Wu, Pingyi Fan, Qiang Fan

Internet of Vehicles (IoV), as the core of intelligent transportation system, enables comprehensive interconnection between vehicles and their surroundings through multiple communication modes, which is significant for autonomous driving and intelligent traffic management. However, with the emergence of new applications, traditional communication technologies face the problems of scarce spectrum resources and high latency. Semantic communication, which focuses on extracting, transmitting, and recovering some useful semantic information from messages, can reduce redundant data transmission, improve spectrum utilization, and provide innovative solutions to communication challenges in the IoV. This paper systematically reviews state of art of semantic communications in the IoV, elaborates the technical background of IoV and semantic communications, and deeply discusses key technologies of semantic communications in IoV, including semantic information extraction, semantic communication architecture, resource allocation and management, and so on. Through specific case studies, it demonstrates that semantic communications can be effectively employed in the scenarios of traffic environment perception and understanding, intelligent driving decision support, IoV service optimization, and intelligent traffic management. Additionally, it analyzes the current challenges and future research directions. This survey reveals that semantic communications has broad application prospects in IoV, but it is necessary to solve the real existing problems by combining advanced technologies to promote its wide application in IoV and contributing to the development of intelligent transportation system.

3. Transformers for Green Semantic Communication: Less Energy, More Semantics

Published: 2023-10-11 Authors: Shubhabrata Mukherjee, Cory Beard, Sejun Song

Semantic communication aims to transmit meaningful and effective information, rather than focusing on individual symbols or bits. This results in benefits like reduced latency, bandwidth usage, and higher throughput compared with traditional communication. However, semantic communication poses significant challenges due to the need for universal metrics to benchmark the joint effects of semantic information loss and practical energy consumption. This research presents a novel multi-objective loss function named “Energy-Optimized Semantic Loss” (EOSL), addressing the challenge of balancing semantic information loss and energy consumption. Through comprehensive experiments on transformer models, including CPU and GPU energy usage, it is demonstrated that EOSL-based encoder model selection can save up to 90% of energy while achieving a 44% improvement in semantic similarity performance during inference in this experiment. This work paves the way for energy-efficient neural network selection and the development of greener semantic communication architectures.

4. From Semantic Communication to Semantic-aware Networking: Model, Architecture, and Open Problems

Published: 2020-12-31 Authors: Guangming Shi, Yong Xiao, Yingyu Li, Xuemei Xie

Existing communication systems are mainly built based on Shannon’s information theory which deliberately ignores the semantic aspects of communication. The recent iteration of wireless technology, the so-called 5G and beyond, promises to support a plethora of services enabled by carefully tailored network capabilities based on the contents, requirements, as well as semantics. This sparkled significant interest in the semantic communication, a novel paradigm that involves the meaning of message into the communication. In this article, we first review the classic semantic communication framework and then summarize key challenges that hinder its popularity. We observe that some semantic communication processes such as semantic detection, knowledge modeling, and coordination, can be resource-consuming and inefficient, especially for the communication between a single source and a destination. We therefore propose a novel architecture based on federated edge intelligence for supporting resource-efficient semantic-aware networking. Our architecture allows each user to offload the computationally intensive semantic encoding and decoding tasks to the edge servers and protect its proprietary model-related information by coordinating via intermediate results. Our simulation result shows that the proposed architecture can reduce the resource consumption and significantly improve the communication efficiency.

5. Semantic-Aware Resource Management for C-V2X Platooning via Multi-Agent Reinforcement Learning

Published: 2024-11-07 Authors: Wenjun Zhang, Qiong Wu, Pingyi Fan, Kezhi Wang, Nan Cheng, Wen Chen, Khaled B. Letaief

Semantic communication transmits the extracted features of information rather than raw data, significantly reducing redundancy, which is crucial for addressing spectrum and energy challenges in 6G networks. In this paper, we introduce semantic communication into a cellular vehicle-to-everything (C-V2X)- based autonomous vehicle platoon system for the first time, aiming to achieve efficient management of communication resources in a dynamic environment. Firstly, we construct a mathematical model for semantic communication in platoon systems, in which the DeepSC model and MU-DeepSC model are used to semantically encode and decode unimodal and multi-modal data, respectively. Then, we propose the quality of experience (QoE) metric based on semantic similarity and semantic rate. Meanwhile, we consider the success rate of semantic information transmission (SRS) metric to ensure the fairness of channel resource allocation. Next, the optimization problem is posed with the aim of maximizing the QoE in vehicle-to-vehicle (V2V) links while improving SRS. To solve this mixed integer nonlinear programming problem (MINLP) and adapt to time-varying channel conditions, the paper proposes a distributed semantic-aware multi-modal resource allocation (SAMRA) algorithm based on multi-agent reinforcement learning (MARL), referred to as SAMRAMARL. The algorithm can dynamically allocate channels and power and determine semantic symbol length based on the contextual importance of the transmitted information, ensuring efficient resource utilization. Finally, extensive simulations have demonstrated that SAMRAMARL outperforms existing methods, achieving significant gains in QoE, SRS, and communication delay in C-V2X platooning scenarios.

6. Reasoning on the Air: An Implicit Semantic Communication Architecture

Published: 2022-02-04 Authors: Yong Xiao, Yingyu Li, Guangming Shi, H. Vincent Poor

Semantic communication is a novel communication paradigm which draws inspiration from human communication focusing on the delivery of the meaning of a message to the intended users. It has attracted significant interest recently due to its potential to improve efficiency and reliability of communication, enhance users’ quality-of-experience (QoE), and achieve smoother cross-protocol/domain communication. Most existing works in semantic communication focus on identifying and transmitting explicit semantic meaning, e.g., labels of objects, that can be directly identified from the source signal. This paper investigates implicit semantic communication in which the hidden information, e.g., implicit causality and reasoning mechanisms of users, that cannot be directly observed from the source signal needs to be transported and delivered to the intended users. We propose a novel implicit semantic communication (iSC) architecture for representing, communicating, and interpreting the implicit semantic meaning. In particular, we first propose a graph-inspired structure to represent implicit meaning of message based on three key components: entity, relation, and reasoning mechanism. We then propose a generative adversarial imitation learning-based reasoning mechanism learning (GAML) solution for the destination user to learn and imitate the reasoning process of the source user. We prove that, by applying GAML, the destination user can accurately imitate the reasoning process of the users to generate reasoning paths that follow the same probability distribution as the expert paths. Numerical results suggest that our proposed architecture can achieve accurate implicit meaning interpretation at the destination user.

7. Knowledge Base Enabled Semantic Communication: A Generative Perspective

Published: 2023-11-21 Authors: Jinke Ren, Zezhong Zhang, Jie Xu, Guanying Chen, Yaping Sun, Ping Zhang, Shuguang Cui

Semantic communication is widely touted as a key technology for propelling the sixth-generation (6G) wireless networks. However, providing effective semantic representation is quite challenging in practice. To address this issue, this article takes a crack at exploiting semantic knowledge base (KB) to usher in a new era of generative semantic communication. Via semantic KB, source messages can be characterized in low-dimensional subspaces without compromising their desired meanings, thus significantly enhancing the communication efficiency. The fundamental principle of semantic KB is first introduced, and a generative semantic communication architecture is developed by presenting three sub-KBs, namely source, task, and channel KBs. Then, the detailed construction approaches for each sub-KB are described, followed by their utilization in terms of semantic coding and transmission. A case study is also provided to showcase the superiority of generative semantic communication over conventional syntactic communication and classical semantic communication. In a nutshell, this article establishes a scientific foundation for the exciting uncharted frontier of generative semantic communication.

8. Rethinking Wireless Communication Security in Semantic Internet of Things

Published: 2022-10-10 Authors: Hongyang Du, Jiacheng Wang, Dusit Niyato, Jiawen Kang, Zehui Xiong, Mohsen Guizani, Dong In Kim

Semantic communication is an important participant in the next generation of wireless communications. Enabled by this novel paradigm, the conventional Internet-of-Things (IoT) is evolving toward the semantic IoT (SIoT) to achieve significant system performance improvements. However, traditional wireless communication security techniques for bit transmission cannot be applied directly to the SIoT that focuses on semantic information transmission. One key reason is the lack of new security performance indicators. Thus, we have to rethink the wireless communication security in the SIoT. As such, in the paper, we analyze and compare classical security techniques, i.e., physical layer security, covert communications, and encryption, from the perspective of semantic information security. We highlight the differences among these security techniques when applied to the SIoT. Novel performance indicators such as semantic secrecy outage probability (for physical layer security techniques) and detection failure probability (for covert communication techniques) are proposed. Considering that semantic communications can raise new security issues, we then review attack and defense methods at the semantic level. Finally, we present several promising directions for future secure SIoT research.

9. On the Computing and Communication Tradeoff in Reasoning-Based Multi-User Semantic Communications

Published: 2024-06-21 Authors: Nitisha Singh, Christo Kurisummoottil Thomas, Walid Saad, Emilio Calvanese Strinati

Semantic communication (SC) is recognized as a promising approach for enabling reliable communication with minimal data transfer while maintaining seamless connectivity for a group of wireless users. Unlocking the advantages of SC for multi-user cases requires revisiting how communication and computing resources are allocated. This reassessment should consider the reasoning abilities of end-users, enabling receiving nodes to fill in missing information or anticipate future events more effectively. Yet, state-of-the-art SC systems primarily focus on resource allocation through compression based on semantic relevance, while overlooking the underlying data generation mechanisms and the tradeoff between communications and computing. Thus, they cannot help prevent a disruption in connectivity. In contrast, in this paper, a novel framework for computing and communication resource allocation is proposed that seeks to demonstrate how SC systems with reasoning capabilities at the end nodes can improve reliability in an end-to-end multi-user wireless system with intermittent communication links. Towards this end, a novel reasoning-aware SC system is proposed for enabling users to utilize their local computing resources to reason the representations when the communication links are unavailable. To optimize communication and computing resource allocation in this system, a noncooperative game is formulated among multiple users whose objective is to maximize the effective semantic information (computed as a product of reliability and semantic information) while controlling the number of semantically relevant links that are disrupted. Simulation results show that the proposed reasoning-aware SC system results in at least a $16.6\%$ enhancement in throughput and a significant improvement in reliability compared to classical communications systems that do not incorporate reasoning.

10. Semantic-Aware Resource Allocation Based on Deep Reinforcement Learning for 5G-V2X HetNets

Published: 2024-06-12 Authors: Zhiyu Shao, Qiong Wu, Pingyi Fan, Nan Cheng, Qiang Fan, Jiangzhou Wang

This letter proposes a semantic-aware resource allocation (SARA) framework with flexible duty cycle (DC) coexistence mechanism (SARADC) for 5G-V2X Heterogeneous Network (HetNets) based on deep reinforcement learning (DRL) proximal policy optimization (PPO). Specifically, we investigate V2X networks within a two-tiered HetNets structure. In response to the needs of high-speed vehicular networking in urban environments, we design a semantic communication system and introduce two resource allocation metrics: high-speed semantic transmission rate (HSR) and semantic spectrum efficiency (HSSE). Our main goal is to maximize HSSE. Additionally, we address the coexistence of vehicular users and WiFi users in 5G New Radio Unlicensed (NR-U) networks. To tackle this complex challenge, we propose a novel approach that jointly optimizes flexible DC coexistence mechanism and the allocation of resources and base stations (BSs). Unlike traditional bit transmission methods, our approach integrates the semantic communication paradigm into the communication system. Experimental results demonstrate that our proposed solution outperforms traditional bit transmission methods with traditional DC coexistence mechanism in terms of HSSE and semantic throughput (ST) for both vehicular and WiFi users.

Architectural & Methodological Insights

The literature from 2020 through 2026 demonstrates a clear evolutionary trajectory in semantic communication (SemCom), transitioning from isolated point-to-point Deep Joint Source-Channel Coding (JSCC) toward holistic, network-wide, and reasoning-aware architectures. Key methodological patterns emerging across these works include:

1. Hybrid Dual-Mode and Knowledge-Base-Assisted Transmission Architectures Early semantic communication relied primarily on end-to-end continuous latent space representations. Modern frameworks increasingly adopt hybrid dual-mode architectures that combine continuous deep feature vectors with discrete Knowledge Base (KB) indexing. By pairing continuous JSCC branches (for high-SNR conditions) with discrete KB-based indexing or memory modules (for low-SNR environments), recent systems mitigate the catastrophic cliff effect typical of pure separation-based or pure continuous systems.

2. Information Bottleneck and Task-Oriented Rate-Distortion Optimization To manage constrained wireless spectrum and heterogeneous hardware, task-oriented semantic extraction is increasingly formalized via the Information Bottleneck (IB) principle. Rather than attempting universal feature compression, architectures apply IB-driven variational objectives, attention mechanisms, and semantic masks to selectively filter task-irrelevant redundancy while preserving perceptual metrics and multi-task capability.

3. Integration with Sensing and Physical Layer Infrastructure (ISAC, IoV, UAVs) Semantic communication is rapidly coalescing with Integrated Sensing and Communication (ISAC) and vehicular networks (IoV/UAVs). Systems now co-optimize multi-modal sensor fusion (e.g., LiDAR, hyperspectral imaging, multi-view cameras) and digital modulation/power allocation across OFDM subcarriers. This creates a closed loop where physical-layer channel conditions directly govern semantic feature granularity and task inference accuracy.

4. LLM-Driven Implicit Semantic Extraction and Reasoning The field is expanding beyond explicit semantic extraction (such as object labels or raw feature maps) toward implicit semantic communication. Leveraging Large Language Models (LLMs), knowledge graphs, and graph-based imitation learning, systems can transmit structured triple representations or implicit causal structures. Receiving nodes utilize generative capabilities and local computing to perform reasoning on the air, allowing recovery of full intent even during intermittent link disruptions.

5. System-Level Orchestration and Compute-Energy Tradeoffs Semantic efficiency is no longer measured solely in saved bandwidth or bit rates. Recent research highlights that neural semantic encoding/decoding incurs non-negligible computational energy overhead. Consequently, the paradigm is shifting toward Semantic-Aware Networking and KB Management and Orchestration (KB-MANO), explicitly balancing communication bandwidth savings against node-level compute/energy expenditure.

Critical Research Gaps

1. Theoretical Modeling and Quantification of High-Order Semantic Impairments While physical channel noise is well understood, high-order semantic impairments—such as semantic ambiguity, knowledge misalignment, and context misinterpretation—lack a standardized, unified mathematical framework. Existing compensation techniques, such as those in “Robust Multimodal Semantic Communications with Semantic Fusion and Compensation” and implicit reasoning models in “Reasoning on the Air: An Implicit Semantic Communication Architecture”, handle specific degradation patterns heuristically. However, a rigorous information-theoretic model defining the capacity limits of channels subject to simultaneous physical and semantic noise remains absent.

2. KB Synchronization and Orchestration Overhead in Non-Stationary Networks Knowledge Bases (KBs) are central to generative and high-compression semantic systems. However, in highly mobile and dynamic environments, continuous knowledge updating and cross-node synchronization introduce substantial signaling overhead and latency. Frameworks proposed in “Collaborative Knowledge Sharing-Empowered Effective Semantic Rate Maximization for Two-Tier Semantic-Bit Communication Networks” and “Semantic Revolution from Communications to Orchestration for 6G: Challenges, Enablers, and Research Directions” highlight this bottleneck, but lack practical, low-latency protocols for real-time KB alignment without consuming the bandwidth saved by semantic compression.

3. Unbounded Energy-Compute Footprint of LLM-Assisted Edge Codecs Integrating LLMs and complex transformer backbones into edge devices drastically improves semantic compression ratios, as shown in “TRILLM: Triple-Based Semantic Communication with LLMS for Robust IIoT Communication”. However, as identified in “Transformers for Green Semantic Communication: Less Energy, More Semantics”, the computational energy consumed during local inference often outweighs the energy saved over the wireless link. Current literature lacks energy-optimized, hardware-aware codec architectures capable of dynamically scaling neural model parameter activation based on instant battery and channel states.

4. Absence of Comprehensive Semantic-Level Physical Security Metrics Traditional security approaches rely on bit-level encryption or physical layer secrecy rates, which are ill-fitted for semantic communications where exact bit recovery is unnecessary. As argued in “Rethinking Wireless Communication Security in Semantic Internet of Things”, current networks lack robust mathematical models to protect against semantic eavesdropping, semantic spoofing, and adversarial perturbations aimed at hijacking downstream task inference.

High-Impact Open Problems

1. Unified Rate-Distortion-Perception-Compute (RDPC) Optimization Framework Develop a generalized theoretical optimization framework that extends classical Rate-Distortion-Perception theory to explicitly incorporate computational energy consumption at both source and destination nodes. The goal is to derive closed-form or bounded Pareto frontiers that determine the optimal operational point between transmitted semantic bit-rate, physical channel distortion, perceptual/task accuracy, and edge-node GPU/CPU energy consumption under time-varying channel conditions.

2. Semantic Physical Layer Security (S-PLS) Against Adversarial Perturbations Formulate a mathematical defense framework against semantic-level adversarial attacks and eavesdropping. This involves designing semantic encryption algorithms and feature-space obfuscation mechanisms grounded in Information Bottleneck constraints. The objective is to maximize the Semantic Secrecy Outage Probability against unauthorized receivers seeking to reconstruct task context, while guaranteeing that legitimate decoders can accurately perform inference with zero degradation in task performance.

3. Dynamic Lightweight KB Synchronization for High-Mobility Edge Networks Design an adaptive, federated Knowledge Base update protocol tailored for dynamic vehicular (C-V2X) and UAV networks. The research should focus on constructing differential semantic update algorithms that isolate and transmit only the delta-knowledge required to realign transmitter and receiver sub-KBs. The protocol must mathematically bound the synchronization latency and guarantee knowledge alignment under rapid topology changes and low SNR conditions.