Related papers: Precision Highway for Ultra Low-Precision Quantiza…

Deep neural networks have been proven effective in a wide range of tasks. However, their high computational and memory costs make them impractical to deploy on resource-constrained devices. To address this issue, quantization schemes have…

We present accumulator-aware quantization (A2Q), a novel weight quantization method designed to train quantized neural networks (QNNs) to avoid overflow when using low-precision accumulators during inference. A2Q introduces a unique…

To deploy deep neural networks on resource-limited devices, quantization has been widely explored. In this work, we study the extremely low-bit networks which have tremendous speed-up, memory saving with quantized activation and weights. We…

Model quantization is a widely used technique to compress and accelerate deep neural network (DNN) inference. Emergent DNN hardware accelerators begin to support mixed precision (1-8 bits) to further improve the computation efficiency,…

Compressing large Neural Networks (NN) by quantizing the parameters, while maintaining the performance is highly desirable due to reduced memory and time complexity. In this work, we cast NN quantization as a discrete labelling problem, and…

Mixed-precision quantization can potentially achieve the optimal tradeoff between performance and compression rate of deep neural networks, and thus, have been widely investigated. However, it lacks a systematic method to determine the…

High resolution compressive channel estimation provides information for vehicle localization when a hybrid mmWave MIMO system is considered. Complexity and memory requirements can, however, become a bottleneck when high accuracy…

Integer quantization of neural networks can be defined as the approximation of the high precision computation of the canonical neural network formulation, using reduced integer precision. It plays a significant role in the efficient…

Binarization is an extreme network compression approach that provides large computational speedups along with energy and memory savings, albeit at significant accuracy costs. We investigate the question of where to binarize inputs at…

Low-bit quantization emerges as one of the most promising compression approaches for deploying deep neural networks on edge devices. Mixed-precision quantization leverages a mixture of bit-widths to unleash the accuracy and efficiency…

Efforts to reduce the numerical precision of computations in deep learning training have yielded systems that aggressively quantize weights and activations, yet employ wide high-precision accumulators for partial sums in inner-product…

Neural image compression has been shown to outperform traditional image codecs in terms of rate-distortion performance. However, quantization introduces errors in the compression process, which can degrade the quality of the compressed…

Although weight and activation quantization is an effective approach for Deep Neural Network (DNN) compression and has a lot of potentials to increase inference speed leveraging bit-operations, there is still a noticeable gap in terms of…

Recognition of overlapped speech has been a highly challenging task to date. State-of-the-art multi-channel speech separation system are becoming increasingly complex and expensive for practical applications. To this end, low-bit neural…

Despite being the cornerstone of deep learning, backpropagation is criticized for its inherent sequentiality, which can limit the scalability of very deep models. Such models faced convergence issues due to vanishing gradient, later…

Efficient model inference is an important and practical issue in the deployment of deep neural network on resource constraint platforms. Network quantization addresses this problem effectively by leveraging low-bit representation and…

We introduce a method to train Quantized Neural Networks (QNNs) --- neural networks with extremely low precision (e.g., 1-bit) weights and activations, at run-time. At train-time the quantized weights and activations are used for computing…

Quantized neural networks typically require smaller memory footprints and lower computation complexity, which is crucial for efficient deployment. However, quantization inevitably leads to a distribution divergence from the original…

Reducing bit-widths of weights, activations, and gradients of a Neural Network can shrink its storage size and memory usage, and also allow for faster training and inference by exploiting bitwise operations. However, previous attempts for…

Quantization has proven effective in high-resolution and large-scale simulations, which benefit from bit-level memory saving. However, identifying a quantization scheme that meets the requirement of both precision and memory efficiency…