We’re delighted to share that our paper, “MANE: A Multi-Path Adaptive Network for Edge Onloading of Deep Neural Networks,” has received the Best Paper Award at the 23rd IEEE International Conference on Ubiquitous Intelligence and Computing (UIC 2026), held in Rende, Italy, September 7–11, 2026.
Led by Sokratis Nikolaidis with collaborators Stylianos I. Venieris, Leonidas Malachias, and Iakovos S. Venieris, MANE tackles a key challenge in distributed inference: how a single edge server can serve many devices at once without being overwhelmed by shared load. Its multi-path tail architecture enables a dynamic accuracy–throughput trade-off at runtime, sustaining over 80% SLO satisfaction where state-of-the-art methods fail completely, while preserving accuracy 6pp higher than on-device alternatives across up to 40 concurrent devices.
Congratulations to Sokratis and the entire team on this well-deserved recognition!
To learn more about the paper click here.