Two-Bit Networks for Deep Learning on Resource-Constrained Embedded Devices

Abstract

With the rapid proliferation of Internet of Things and intelligent edge devices, there is an increasing need for implementing machine learning algorithms, including deep learning, on resource-constrained mobile embedded devices with limited memory and computation power. Typical large Convolutional Neural Networks (CNNs) need large amounts of memory and computational power, and cannot be deployed on embedded devices efficiently. We present Two-Bit Networks (TBNs) for model compression of CNNs with edge weights constrained to (-2, -1, 1, 2), which can be encoded with two bits. Our approach can reduce the memory usage and improve computational efficiency significantly while achieving good performance in terms of classification accuracy, thus representing a reasonable tradeoff between model size and performance.

References in corpus (1)

• Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1

Cited by in corpus (10)

• FermiNets: Learning generative machines to generate efficient neural networks via generative synthesis
• Fog-Assisted wIoT: A Smart Fog Gateway for End-to-End Analytics in Wearable Internet of Things
• SquishedNets: Squishing SqueezeNet further for edge device scenarios via deep evolutionary synthesis
• JSDoop and TensorFlow.js: Volunteer Distributed Web Browser-Based Neural Network Training
• AttendNets: Tiny Deep Image Recognition Neural Networks for the Edge via Visual Attention Condensers
• BitNet: Bit-Regularized Deep Neural Networks
• EdgeSegNet: A Compact Network for Semantic Segmentation
• AttendSeg: A Tiny Attention Condenser Neural Network for Semantic Segmentation on the Edge
• Model compression for faster structural separation of macromolecules captured by Cellular Electron Cryo-Tomography
• An Overview of Datatype Quantization Techniques for Convolutional Neural Networks