Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2504.21136 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910922597990400 |
|---|---|
| author | Ramanujam, Murali Dai, Yinwei Jamieson, Kyle Netravali, Ravi |
| author_facet | Ramanujam, Murali Dai, Yinwei Jamieson, Kyle Netravali, Ravi |
| contents | Continually retraining models has emerged as a primary technique to enable high-accuracy video analytics on edge devices. Yet, existing systems employ such adaptation by relying on the spare compute resources that traditional (memory-constrained) edge servers afford. In contrast, mobile edge devices such as drones and dashcams offer a fundamentally different resource profile: weak(er) compute with abundant unified memory pools. We present Legilimens, a continuous learning system for the mobile edge's System-on-Chip GPUs. Our driving insight is that visually distinct scenes that require retraining exhibit substantial overlap in model embeddings; if captured into a base model on device memory, specializing to each new scene can become lightweight, requiring very few samples. To practically realize this approach, Legilimens presents new, compute-efficient techniques to (1) select high-utility data samples for retraining specialized models, (2) update the base model without complete retraining, and (3) time-share compute resources between retraining and live inference for maximal accuracy. Across diverse workloads, Legilimens lowers retraining costs by 2.8-10x compared to existing systems, resulting in 18-45% higher accuracies. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_21136 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Legilimens: Performant Video Analytics on the System-on-Chip Edge Ramanujam, Murali Dai, Yinwei Jamieson, Kyle Netravali, Ravi Computer Vision and Pattern Recognition Machine Learning Continually retraining models has emerged as a primary technique to enable high-accuracy video analytics on edge devices. Yet, existing systems employ such adaptation by relying on the spare compute resources that traditional (memory-constrained) edge servers afford. In contrast, mobile edge devices such as drones and dashcams offer a fundamentally different resource profile: weak(er) compute with abundant unified memory pools. We present Legilimens, a continuous learning system for the mobile edge's System-on-Chip GPUs. Our driving insight is that visually distinct scenes that require retraining exhibit substantial overlap in model embeddings; if captured into a base model on device memory, specializing to each new scene can become lightweight, requiring very few samples. To practically realize this approach, Legilimens presents new, compute-efficient techniques to (1) select high-utility data samples for retraining specialized models, (2) update the base model without complete retraining, and (3) time-share compute resources between retraining and live inference for maximal accuracy. Across diverse workloads, Legilimens lowers retraining costs by 2.8-10x compared to existing systems, resulting in 18-45% higher accuracies. |
| title | Legilimens: Performant Video Analytics on the System-on-Chip Edge |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2504.21136 |