.png){kind=icon}
.png){kind=icon}
Return
Fast Wireless Foundation Models With Early-Exits
.png){kind=icon}
PRE
AI
O
H
DOI：10.1109/lcomm.2026.3710677.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
While wireless foundation models (FMs) are demonstrating strong potential to enable AI-Native 6G networks, their high computational cost remains a critical barrier to deployment. The large computational cost stems from the rigid, full-depth execution of the FM backbone for every task − a process we show is not only inefficient but can also degrade performance on unseen out-of-distribution (OOD) tasks. In this letter, we propose a novel early-exit FM framework that attaches lightweight, per-task heads, at the most appropriate exit-stage of a frozen wireless FM encoder, enabling variable-depth inference tailored to each task’s preferred representation depth. Our results demonstrate that these intermediate-layer features not only speed-up inference significantly (up to 93% fewer FLOPs), but also provide more transferable representations that exceed the full encoder accuracy on unseen tasks. We further demonstrate that a simple fixed-exit strategy per task is more effective than traditional early-exiting policies that route different samples to different exits based on their perceived difficulty levels.
Keywords:
Foundation models
early exiting
OOD generalization
MIMO systems
Journal
IF:
4.4
Papers:
1.2W
Citations:
2.2W
Organization
Cited Papers