[Paper Review] Energy Efficient Semantic Communication over Wireless Networks with Rate Splitting

Zhaohui Yang, Mingzhe Chen+2 more|arXiv (Cornell University)|Jan 5, 2023

Energy Harvesting in Wireless Networks4 citations

TL;DR

This paper proposes an energy-efficient semantic communication framework for wireless networks using rate splitting (RSMA), where a base station extracts multi-level semantic information from large data and transmits compact semantic messages via common and private streams. By jointly optimizing semantic extraction ratio and computation frequency, the scheme minimizes total energy under latency and power constraints, achieving near-optimal performance with lower energy than FDMA, NOMA, and SDMA, especially at high transmit power.

ABSTRACT

In this paper, the problem of wireless resource allocation and semantic information extraction for energy efficient semantic communications over wireless networks with rate splitting is investigated. In the considered model, a base station (BS) first extracts semantic information from its large-scale data, and then transmits the small-sized semantic information to each user which recovers the original data based on its local common knowledge. At the BS side, the probability graph is used to extract multi-level semantic information. In the downlink transmission, a rate splitting scheme is adopted, while the private small-sized semantic information is transmitted through private message and the common knowledge is transmitted through common message. Due to limited wireless resource, both computation energy and transmission energy are considered. This joint computation and communication problem is formulated as an optimization problem aiming to minimize the total communication and computation energy consumption of the network under computation, latency, and transmit power constraints. To solve this problem, an alternating algorithm is proposed where the closed-form solutions for semantic information extraction ratio and computation frequency are obtained at each step. Numerical results verify the effectiveness of the proposed algorithm.

Motivation & Objective

• To address the energy efficiency challenge in next-generation wireless networks supporting high-data-rate, low-latency applications like metaverse and digital twins.
• To design a joint computation and communication resource allocation framework that minimizes total energy consumption under latency, power, and semantic accuracy constraints.
• To enable semantic communication by extracting multi-level semantic information from large-scale data at the base station and transmitting compact semantic messages via rate splitting.
• To optimize both computation (at users) and transmission (over wireless links) energy by jointly controlling semantic extraction ratio and computation frequency.

Proposed method

• Uses a probability graph to extract multi-level semantic information from large-scale data at the base station.
• Applies rate splitting multiple access (RSMA) in downlink transmission, splitting messages into common (shared knowledge) and private (user-specific) parts.
• Models the system with joint computation and communication energy minimization, incorporating transmit power, latency, and computation capacity constraints.
• Develops an alternating optimization algorithm that derives closed-form solutions for semantic extraction ratio and computation frequency at each iteration.
• Solves the non-convex optimization problem by iteratively updating variables while maintaining feasibility under all constraints.
• Employs an exhaustive search baseline (EXH-RSMA) to validate the near-optimal performance of the proposed algorithm.

Figure 1: Illustration of the considered SWC network with rate splitting.

Experimental results

Research questions

• RQ1 How can semantic information be efficiently extracted from large-scale data to reduce transmission overhead in wireless networks?
• RQ2 What is the optimal trade-off between computation and communication energy in a semantic communication system with limited wireless resources?
• RQ3 How does rate splitting improve energy efficiency compared to conventional multiple access schemes like FDMA, NOMA, and SDMA?
• RQ4 What is the impact of transmit power, bandwidth, and user computation capacity on the total energy consumption of the semantic communication system?
• RQ5 Can a joint optimization of semantic extraction and computation frequency achieve near-global optimality with low-complexity algorithms?

Key findings

• The proposed RSMA-based semantic communication scheme achieves significantly lower total energy consumption than FDMA, NOMA, and SDMA, especially at high transmit power levels.
• Total energy decreases with increasing maximum transmit power due to reduced transmission time, allowing more time for computation and lower power consumption.
• Higher system bandwidth reduces total energy by shortening transmission time and enabling longer computation durations.
• The proposed algorithm achieves near-optimal performance, closely matching the exhaustive search baseline (EXH-RSMA), demonstrating its effectiveness and efficiency.
• The total energy consumption grows more slowly with data size in RSMA compared to NOMA and FDMA, indicating greater robustness.
• For low to moderate computation capacity, increasing maximum computation capacity sharply reduces total energy; beyond a threshold, further increases have negligible impact, indicating saturation in energy savings.

Figure 2: An example of the multi-level semantic information extraction.