Performance Guaranteed Network Acceleration via High-Order Residual Quantization
This work addresses network acceleration for deep learning applications, offering a novel method to reduce accuracy loss compared to previous binarization approaches.
The paper tackles the problem of accuracy loss in network acceleration through input binarization by proposing a high-order binarization scheme that recursively performs residual quantization, achieving more accurate approximation while maintaining binary operation advantages. Experimental results show it yields great recognition accuracy with acceleration.
Input binarization has shown to be an effective way for network acceleration. However, previous binarization scheme could be regarded as simple pixel-wise thresholding operations (i.e., order-one approximation) and suffers a big accuracy loss. In this paper, we propose a highorder binarization scheme, which achieves more accurate approximation while still possesses the advantage of binary operation. In particular, the proposed scheme recursively performs residual quantization and yields a series of binary input images with decreasing magnitude scales. Accordingly, we propose high-order binary filtering and gradient propagation operations for both forward and backward computations. Theoretical analysis shows approximation error guarantee property of proposed method. Extensive experimental results demonstrate that the proposed scheme yields great recognition accuracy while being accelerated.