Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Inference-Driven Uplink for 6G: Architecture, Principles, and Challenges

Chunmei Xu, Zhi Ding, Yi Ma, Rahim Tafazolli, Peiying Zhu

TL;DR

InferCom tackles the 6G uplink bottleneck by moving heavy semantic inference from the constrained transmitter to a GenAI-based receiver, guided by the information bottleneck framework. It introduces a lightweight, task-agnostic transmitter, an inference-capable receiver, and a QoE-aware feedback loop to preserve task relevance under severe channel degradation. Case studies show InferCom delivering superior semantic QoE and substantial transmitter-computation savings, while reducing retransmissions compared with 5G NR and Deep-JSCC. The work highlights practical challenges in metric design, cross-layer protocol design, energy-aware model selection, multi-user access, and security, outlining a research roadmap toward human-centric, intelligent 6G networks.

Abstract

Next-generation wireless networks (6G) face a critical uplink challenge arising from stringent device-side resource constraints and the growing demand for intelligence services. This article introduces InferCom, an inference-driven communication architecture designed to enable robust 6G uplink transmission under low signal-to-noise (SNR) conditions. InferCom adopts a compute-asymmetric architecture, featuring a lightweight transmitter and an inference-capable receiver empowered by generative artificial intelligence (GenAI) models, together with a quality-of-experience (QoE)-aware retransmission mechanism. Grounded in the information bottleneck (IB) theory, InferCom redefines uplink communications through task-agnostic compression, inference-driven reconstruction, error-distribution channel coding, and QoE-aware feedback. The case study demonstrates that InferCom outperforms conventional 5G NR and Deep- JSCC in terms of transmitter-side computational complexity, required SNRs and retransmission efficiency. Finally, we outline key challenges and research directions toward making InferCom a practical enabler of human-centric, intelligent and sustainable wireless networks.

Inference-Driven Uplink for 6G: Architecture, Principles, and Challenges

TL;DR

InferCom tackles the 6G uplink bottleneck by moving heavy semantic inference from the constrained transmitter to a GenAI-based receiver, guided by the information bottleneck framework. It introduces a lightweight, task-agnostic transmitter, an inference-capable receiver, and a QoE-aware feedback loop to preserve task relevance under severe channel degradation. Case studies show InferCom delivering superior semantic QoE and substantial transmitter-computation savings, while reducing retransmissions compared with 5G NR and Deep-JSCC. The work highlights practical challenges in metric design, cross-layer protocol design, energy-aware model selection, multi-user access, and security, outlining a research roadmap toward human-centric, intelligent 6G networks.

Abstract

Next-generation wireless networks (6G) face a critical uplink challenge arising from stringent device-side resource constraints and the growing demand for intelligence services. This article introduces InferCom, an inference-driven communication architecture designed to enable robust 6G uplink transmission under low signal-to-noise (SNR) conditions. InferCom adopts a compute-asymmetric architecture, featuring a lightweight transmitter and an inference-capable receiver empowered by generative artificial intelligence (GenAI) models, together with a quality-of-experience (QoE)-aware retransmission mechanism. Grounded in the information bottleneck (IB) theory, InferCom redefines uplink communications through task-agnostic compression, inference-driven reconstruction, error-distribution channel coding, and QoE-aware feedback. The case study demonstrates that InferCom outperforms conventional 5G NR and Deep- JSCC in terms of transmitter-side computational complexity, required SNRs and retransmission efficiency. Finally, we outline key challenges and research directions toward making InferCom a practical enabler of human-centric, intelligent and sustainable wireless networks.

Paper Structure

This paper contains 24 sections, 4 figures.

Table of Contents

  1. Introduction
  2. InferCom Architecture
  3. Inference-Oriented Transmitter
  4. Inference-Capable Receiver
  5. QoE-Aware Retransmission Mechanism
  6. Summary of Architectural Advantages
  7. Theory and Design Principles of InferCom
  8. Simple, Task-Agnostic Compression
  9. Inference-Driven Reconstruction
  10. Error-Distributing Channel Code
  11. QoE-Aware Retransmissions
  12. Summary
  13. Case Study and Illustrative Results
  14. Case Study Setup
  15. Illustrative Results
  16. ...and 9 more sections

Figures (4)

  • Figure 1: The proposed InferCom architecture for 6G uplink.
  • Figure 2: Theory and design principles of InferCom.
  • Figure 3: Illustrative results of InferCom, 5G NR, and Deep-JSCC under low-SNR conditions, where the compression rate is set to around $0.09$ and Deep-JSCC is trained under $\mathrm{SNR}=0$ dB.
  • Figure 4: Performance comparison: (a) Transmitter-side computational complexity, where the standardized quasi-cyclic (QC) LDPC code using base graph 1 for 5B NR baseline. (b) SNR gains across different compression rates. (d) Retransmission rates across low-SNR range.