DNN Feature Map Compression using Learned Representation over GF(2)

Abstract

In this paper, we introduce a method to compress intermediate feature maps of deep neural networks (DNNs) to decrease memory storage and bandwidth requirements during inference. Unlike previous works, the proposed method is based on converting fixed-point activations into vectors over the smallest GF(2) finite field followed by nonlinear dimensionality reduction (NDR) layers embedded into a DNN. Such an end-to-end learned representation finds more compact feature maps by exploiting quantization redundancies within the fixed-point activations along the channel or spatial dimensions. We apply the proposed network architectures derived from modified SqueezeNet and MobileNetV2 to the tasks of ImageNet classification and PASCAL VOC object detection. Compared to prior approaches, the conducted experiments show a factor of 2 decrease in memory requirements with minor degradation in accuracy while adding only bitwise computations.

arXiv commentCEFRL2018

References in corpus (10)

• SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation
• DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
• Binarized Neural Networks
• Convolutional Neural Networks using Logarithmic Data Representation
• Training deep neural networks with low precision multiplications
• XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks
• Hardware-oriented Approximation of Convolutional Neural Networks
• Sigma Delta Quantized Networks