Low-Cost On-device Partial Domain Adaptation (LoCO-PDA): Enabling efficient CNN retraining on edge devices
This work addresses the challenge of efficient partial domain adaptation for edge devices, enabling on-device retraining to mitigate performance degradation in real-world deployments.
The paper tackles the problem of adapting convolutional neural networks (CNNs) to new data distributions on edge devices, proposing LoCO-PDA, which improves classification accuracy by 3.04 percentage points on average while reducing retraining memory consumption by up to 15.1x and improving inference latency by 2.07x on an NVIDIA Jetson TX2.
With the increased deployment of Convolutional Neural Networks (CNNs) on edge devices, the uncertainty of the observed data distribution upon deployment has led researchers to to utilise large and extensive datasets such as ILSVRC'12 to train CNNs. Consequently, it is likely that the observed data distribution upon deployment is a subset of the training data distribution. In such cases, not adapting a network to the observed data distribution can cause performance degradation due to negative transfer and alleviating this is the focus of Partial Domain Adaptation (PDA). Current works targeting PDA do not focus on performing the domain adaptation on an edge device, adapting to a changing target distribution or reducing the cost of deploying the adapted network. This work proposes a novel PDA methodology that targets all of these directions and opens avenues for on-device PDA. LoCO-PDA adapts a deployed network to the observed data distribution by enabling it to be retrained on an edge device. Across subsets of the ILSVRC12 dataset, LoCO-PDA improves classification accuracy by 3.04pp on average while achieving up to 15.1x reduction in retraining memory consumption and 2.07x improvement in inference latency on the NVIDIA Jetson TX2. The work is open-sourced at \emph{link removed for anonymity}.