The quantized neural networks (QNNs) can be useful for neural network acceleration and compression, but during the training process they pose a challenge: how to propagate the gradient of loss function through the graph flow with a derivative of 0 almost everywhere. In response to this non-differentiable situation, we propose a novel Asymptotic-Quantized Estimator (AQE) to estimate the gradient. In particular, during back-propagation, the graph that relates inputs to output remains smoothness and differentiability. At the end of training, the weights and activations have been quantized to low-precision because of the asymptotic behaviour of AQE. Meanwhile, we propose a M-bit Inputs and N-bit Weights Network (MINW-Net) trained by AQE, a quantized neural network with 1-3 bits weights and activations. In the inference phase, we can use XNOR or SHIFT operations instead of convolution operations to accelerate the MINW-Net. Our experiments on CIFAR datasets demonstrate that our AQE is well defined, and the QNNs with AQE perform better than that with Straight-Through Estimator (STE). For example, in the case of the same ConvNet that has 1-bit weights and activations, our MINW-Net with AQE can achieve a prediction accuracy 1.5\% higher than the Binarized Neural Network (BNN) with STE. The MINW-Net, which is trained from scratch by AQE, can achieve comparable classification accuracy as 32-bit counterparts on CIFAR test sets. Extensive experimental results on ImageNet dataset show great superiority of the proposed AQE and our MINW-Net achieves comparable results with other state-of-the-art QNNs.
Deep neural network quantization with adaptive bitwidths has gained increasing attention due to the ease of model deployment on various platforms with different resource budgets. In this paper, we propose a meta-learning approach to achieve this goal. Specifically, we propose MEBQAT, a simple yet effective way of bitwidth-adaptive quantization aware training (QAT) where meta-learning is effectively combined with QAT by redefining meta-learning tasks to incorporate bitwidths. After being deployed on a platform, MEBQAT allows the (meta-)trained model to be quantized to any candidate bitwidth the…
University of Chinese Academy of Sciences（中国科学院大学）；Megvii Inc.（旷视科技有限公司）
Quantized Neural Networks (QNNs), which use low bitwidth numbers for representing parameters and performing computations, have been proposed to reduce the computation complexity, storage size and memory usage. In QNNs, parameters and activations are uniformly quantized, such that the multiplications and additions can be accelerated by bitwise operations. However, distributions of parameters in Neural Networks are often imbalanced, such that the uniform quantization determined from extremal values may under utilize available bitwidth. In this paper, we propose a novel quantization method that c…
Eindhoven University of Technology（埃因霍温理工大学）；Philips Innovation Services（飞利浦创新服务）
Convolutional Neural Networks (CNNs) have proven to be a powerful state-of-the-art method for image classification tasks. One drawback however is the high computational complexity and high memory consumption of CNNs which makes them unfeasible for execution on embedded platforms which are constrained on physical resources needed to support CNNs. Quantization has often been used to efficiently optimize CNNs for memory and computational complexity at the cost of a loss of prediction accuracy. We therefore propose a method to optimally quantize the weights, biases and activations of each layer of…
Monash University（莫纳什大学）；The University of Adelaide（阿德莱德大学）；South China University of Technology（华南理工大学）
This paper tackles the problem of training a deep convolutional neural network of both low-bitwidth weights and activations. Optimizing a low-precision network is very challenging due to the non-differentiability of the quantizer, which may result in substantial accuracy loss. To address this, we propose three practical approaches, including (i) progressive quantization; (ii) stochastic precision; and (iii) joint knowledge distillation to improve the network training. First, for progressive quantization, we propose two schemes to progressively find good local minima. Specifically, we propose t…
As the applications of deep learning models on edge devices increase at an accelerating pace, fast adaptation to various scenarios with varying resource constraints has become a crucial aspect of model deployment. As a result, model optimization strategies with adaptive configuration are becoming increasingly popular. While single-shot quantized neural architecture search enjoys flexibility in both model architecture and quantization policy, the combined search space comes with many challenges, including instability when training the weight-sharing supernet and difficulty in navigating the exp…
Quantizing neural networks to low-bitwidth is important for model deployment on resource-limited edge hardware. Although a quantized network has a smaller model size and memory footprint, it is fragile to adversarial attacks. However, few methods study the robustness and training efficiency of quantized networks. To this end, we propose a new method by recasting robust quantization as an online domain generalization problem, termed ODG-Q, which generates diverse adversarial data at a low cost during training. ODG-Q consistently outperforms existing works against various adversarial attacks. Fo…
Deep quantization of neural networks (below eight bits) offers significant promise in reducing their compute and storage cost. Albeit alluring, without special techniques for training and optimization, deep quantization results in significant accuracy loss. To further mitigate this loss, we propose a novel sinusoidal regularization, called SinReQ1, for deep quantized training. SinReQ adds a periodic term to the original objective function of the underlying training algorithm. SinReQ exploits the periodicity, differentiability, and the desired convexity profile in sinusoidal functions to automa…
Model quantization is a widely used technique to compress and accelerate deep neural network (DNN) inference, especially when deploying to edge or IoT devices with limited computation capacity and power consumption budget. The uniform bit width quantization across all the layers is usually sub-optimal and the exploration of hybrid quantization for different layers is vital for efficient deep compression. In this paper, we employ the meta learning method to automatically realize low-bit hybrid quantization of neural networks. A MetaQuantNet, together with a Quantization function, are trained to…
We present PrecisionBatching, a quantized inference algorithm for speeding up neural network execution on traditional hardware platforms at low bitwidths without the need for retraining or recalibration. PrecisionBatching decomposes a neural network into individual bitlayers and accumulates them using fast 1-bit operations while maintaining activations in full precision. PrecisionBatching not only facilitates quantized inference at low bitwidths (< 8 bits) without the need for retraining/recalibration, but also 1) enables traditional hardware platforms the ability to realize inference speedups…
We propose DoReFa-Net, a method to train convolutional neural networks that have low bitwidth weights and activations using low bitwidth parameter gradients. In particular, during backward pass, parameter gradients are stochastically quantized to low bitwidth numbers before being propagated to convolutional layers. As convolutions during forward/backward passes can now operate on low bitwidth weights and activations/gradients respectively, DoReFa-Net can use bit convolution kernels to accelerate both training and inference. Moreover, as bit convolutions can be efficiently implemented on CPU, F…
Fully quantized training (FQT), which uses low-bitwidth hardware by quantizing the activations, weights, and gradients of a neural network model, is a promising approach to accelerate the training of deep neural networks. One major challenge with FQT is the lack of theoretical understanding, in particular of how gradient quantization impacts convergence properties. In this paper, we address this problem by presenting a statistical framework for analyzing FQT algorithms. We view the quantized gradient of FQT as a stochastic estimator of its full precision counterpart, a procedure known as quant…
Performance Optimization for Variable Bitwidth Federated Learning in Wireless Networks
Beijing University of Posts and Telecommunications（北京邮电大学）；University of Miami（迈阿密大学）；Purdue University（普渡大学）；Virginia Tech（弗吉尼亚理工大学）；The Chinese University of Hong Kong（香港中文大学）；Shenzhen Research Institute of Big Data（深圳大数据研究院）；Peng Cheng Laboratory（鹏城实验室）
Recent advance of large scale similarity search involves using deeply learned representations to improve the search accuracy and use vector quantization methods to increase the search speed. However, how to learn deep representations that strongly preserve similarities between data pairs and can be accurately quantized via vector quantization remains a challenging task. Existing methods simply leverage quantization loss and similarity loss, which result in unexpectedly biased back-propagating gradients and affect the search performances. To this end, we propose a novel gradient snapping layer …
University of California San Diego（加州大学圣地亚哥分校）；CNRS（法国国家科学研究中心）；Université de Bordeaux（波尔多大学）；Bordeaux INP（波尔多国立理工学院）
As deep neural networks make their ways into different domains, their compute efficiency is becoming a first-order constraint. Deep quantization, which reduces the bitwidth of the operations (below 8 bits), offers a unique opportunity as it can reduce both the storage and compute requirements of the network super-linearly. However, if not employed with diligence, this can lead to significant accuracy loss. Due to the strong inter-dependence between layers and exhibiting different characteristics across the same network, choosing an optimal bitwidth per layer granularity is not a straight forwa…
CEA Tech（法国原子能与替代能源委员会技术中心）；Mines Saint-Etienne（圣埃蒂安矿业学院）
As the will to deploy neural networks models on embedded systems grows, and considering the related memory footprint and energy consumption issues, finding lighter solutions to store neural networks such as weight quantization and more efficient inference methods become major research topics. Parallel to that, adversarial machine learning has risen recently with an impressive and significant attention, unveiling some critical flaws of machine learning models, especially neural networks. In particular, perturbed inputs called adversarial examples have been shown to fool a model into making inco…
QLLM: Accurate and Efficient Low-Bitwidth Quantization for Large Language Models
ZIP Lab, Monash University（莫纳什大学ZIP实验室）；SenseTime Research（商汤科技研究院）；Beihang University（北京航空航天大学）；School of Computer and Communication Sciences, EPFL（洛桑联邦理工学院计算机与通信科学学院）；University of Science and Technology Beijing（北京科技大学）
NestQuant: Post-Training Integer-Nesting Quantization for On-Device DNN
Key Laboratory of Big Data & Artificial Intelligence in Transportation (Beijing Jiaotong University), Ministry of Education（大数据与人工智能在交通中的关键实验室（北京交通大学，教育部））；School of Computer Science and Technology, Beijing Jiaotong University（计算机科学与技术学院，北京交通大学）；School of Artificial Intelligence, Beijing Normal University（人工智能学院，北京师范大学）；Department of Computer Science, City University of Hong Kong（计算机科学系，香港城市大学）
SenseTime Research（商汤科技研究院）；University of Electronic Science and Technology of China（电子科技大学）
We study the challenging task of neural network quantization without end-to-end retraining, called Post-training Quantization (PTQ). PTQ usually requires a small subset of training data but produces less powerful quantized models than Quantization-Aware Training (QAT). In this work, we propose a novel PTQ framework, dubbed BRECQ, which pushes the limits of bitwidth in PTQ down to INT2 for the first time. BRECQ leverages the basic building blocks in neural networks and reconstructs them one-by-one. In a comprehensive theoretical study of the second-order error, we show that BRECQ achieves a goo…
University of Electronic Science and Technology of China（电子科技大学）；Harvard University（哈佛大学）；The Chinese University of Hong Kong（香港中文大学）；National University of Singapore（新加坡国立大学）
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 first bring up three omitted issues in extremely low-bit networks: the squashing range of quantized values; the gradient vanishing during backpropagation and the unexploited hardware acceleration of ternary networks. By reparameterizing quantized activation and weights vector with full precision scale and offset for fixed ternary vector, we decouple the ran…
Outliers in weights and activations pose a key challenge for fixed-point quantization of neural networks. While they can be addressed by fine-tuning, this is not practical for ML service providers (e.g., Google or Microsoft) who often receive customer models without training data. Specialized hardware for handling activation outliers can enable low-precision neural networks, but at the cost of nontrivial area overhead. We instead propose overwrite quantization (OverQ), a lightweight hardware technique that opportunistically increases bitwidth for activation outliers by overwriting nearby zeros…
Network quantization is one of the most hardware friendly techniques to enable the deployment of convolutional neural networks (CNNs) on low-power mobile devices. Recent network quantization techniques quantize each weight kernel in a convolutional layer independently for higher inference accuracy, since the weight kernels in a layer exhibit different variances and hence have different amounts of redundancy. The quantization bitwidth or bit number (QBN) directly decides the inference accuracy, latency, energy and hardware overhead. To effectively reduce the redundancy and accelerate CNN infere…
Universita’ di Bologna（博洛尼亚大学）；ETH Zurich（苏黎世联邦理工学院）
This paper presents a novel end-to-end methodology for enabling the deployment of low-error deep networks on microcontrollers. To fit the memory and computational limitations of resource-constrained edge-devices, we exploit mixed low-bitwidth compression, featuring 8, 4 or 2-bit uniform quantization, and we model the inference graph with integer-only operations. Our approach aims at determining the minimum bit precision of every activation and weight tensor given the memory constraints of a device. This is achieved through a rule-based iterative procedure, which cuts the number of bits of the …
The Straight-Through Estimator (STE) is widely used for back-propagating gradients through the quantization function, but the STE technique lacks a complete theoretical understanding. We propose an alternative methodology called alpha-blending (AB), which quantizes neural networks to low-precision using stochastic gradient descent (SGD). Our method (AB) avoids STE approximation by replacing the quantized weight in the loss function by an affine combination of the quantized weight w_q and the corresponding full-precision weight w with non-trainable scalar coefficient $α$ and $1-α$. During train…
