Data-Side Efficiencies for Lightweight Convolutional Neural Networks
This work addresses the need for efficient model selection in resource-constrained applications like robotics, though it is incremental as it builds on existing lightweight network and metric learning methods.
The paper tackles the problem of selecting lightweight convolutional neural networks for visual tasks by analyzing how data attributes like number of classes and image resolution affect model efficiency, showing that their metrics reduce computation by 30x compared to full inference tests and achieve a 66% computation reduction with a 3.5% accuracy gain in a robot path planning application.
We examine how the choice of data-side attributes for two important visual tasks of image classification and object detection can aid in the choice or design of lightweight convolutional neural networks. We show by experimentation how four data attributes - number of classes, object color, image resolution, and object scale affect neural network model size and efficiency. Intra- and inter-class similarity metrics, based on metric learning, are defined to guide the evaluation of these attributes toward achieving lightweight models. Evaluations made using these metrics are shown to require 30x less computation than running full inference tests. We provide, as an example, applying the metrics and methods to choose a lightweight model for a robot path planning application and achieve computation reduction of 66% and accuracy gain of 3.5% over the pre-method model.