Abstract | ||
---|---|---|
ABSTRACTOn-device machine learning is becoming a reality thanks to the availability of powerful hardware and model compression techniques. Typically, these models are pretrained on large GPU clusters and have enough parameters to generalise across a wide variety of inputs. In this work, we observe that a much smaller, personalised model can be employed to fit a specific scenario, resulting in both higher accuracy and faster execution. Nevertheless, on-device training is extremely challenging, imposing excessive computational and memory requirements even for flagship smartphones. At the same time, on-device data availability might be limited and samples are most frequently unlabelled. To this end, we introduce PersEPhonEE, a framework that attaches early exits on the model and personalises them on-device. These allow the model to progressively bypass a larger part of the computation as more personalised data become available. Moreover, we introduce an efficient on-device algorithm that trains the early exits in a semi-supervised manner at a fraction of the whole network's personalisation time. Results show that PersEPhonEE boosts accuracy by up to 15.9% while dropping the training cost by up to 2.2× and inference latency by 2.2-3.2× on average for the same accuracy, depending on the availability of labels on-device. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1145/3446382.3448359 | HOTMOBILE |
DocType | Citations | PageRank |
Conference | 2 | 0.40 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ilias Leontiadis | 1 | 761 | 44.38 |
Stefanos Laskaridis | 2 | 18 | 3.21 |
Stylianos I. Venieris | 3 | 106 | 12.98 |
Nicholas D. Lane | 4 | 4247 | 248.15 |