🎉 Paper Accepted at MobiSys 2026

Feb 28, 2026·
Admin
· 1 min read

Great news! Our paper has been conditionally accepted at ACM MobiSys 2026!

Paper Details

Title: KVSwap: Disk-aware KV Cache Offloading for Long-Context On-device Inference

Authors:

  • Huawei Zhang (University of Leeds)
  • Chunwei Xia (University of Leeds)
  • Zheng Wang (University of Leeds)

Paper ID: #610

Conference Information

  • Conference: ACM MobiSys 2026 (24th ACM International Conference on Mobile Systems, Applications, and Services)
  • Date: June 21-25, 2026
  • Location: Cambridge, UK
  • Acceptance Rate: 68/304 submissions (22.4%)

About the Work

This paper addresses an important challenge in deploying large language models on mobile and edge devices with limited memory. We propose KVSwap, a novel approach for disk-aware KV cache offloading that enables efficient long-context inference on resource-constrained devices.

MobiSys is a premier venue for innovative research on mobile computing, applications, and services. We’re excited to present this work and contribute to advancing on-device AI capabilities!

The reviews and shepherding instructions will be available soon. Looking forward to the shepherding process and the conference!

Publication MobiSys On-Device Inference KV Cache