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 forward. As such, deep quantization opens a large hyper-parameter space, the exploration of which is a major challenge. We propose a novel sinusoidal regularization, called SINAREQ, for deep quantized training. Leveraging the sinusoidal properties, we seek to learn multiple quantization parameterization in conjunction during gradient-based training process. Specifically, we learn (i) a per-layer quantization bitwidth along with (ii) a scale factor through learning the period of the sinusoidal function. At the same time, we exploit the periodicity, differentiability, and the local convexity profile in sinusoidal functions to automatically propel (iii) network weights towards values quantized at levels that are jointly determined. We show how SINAREQ balance compute efficiency and accuracy, and provide a heterogeneous bitwidth assignment for quantization of a large variety of deep networks (AlexNet, CIFAR-10, MobileNet, ResNet-18, ResNet-20, SVHN, and VGG-11) that virtually preserves the accuracy. Furthermore, we carry out experimentation using fixed homogenous bitwidths with 3- to 5-bit assignment and show the versatility of SINAREQ in enhancing quantized training algorithms (DoReFa and WRPN) with about 4.8% accuracy improvements on average, and then outperforming multiple state-of-the-art techniques.
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…
Qualcomm AI Research（高通人工智能研究院）；Qualcomm Technologies Netherlands B.V.（荷兰高通技术有限公司）；University of Oxford（牛津大学）
We analyze the effect of quantizing weights and activations of neural networks on their loss and derive a simple regularization scheme that improves robustness against post-training quantization. By training quantization-ready networks, our approach enables storing a single set of weights that can be quantized on-demand to different bit-widths as energy and memory requirements of the application change. Unlike quantization-aware training using the straight-through estimator that only targets a specific bit-width and requires access to training data and pipeline, our regularization-based method…
KTH Royal Institute of Technology（瑞典皇家理工学院）；Research Institutes of Sweden（瑞典研究机构）
Autoencoders and their variations provide unsupervised models for learning low-dimensional representations for downstream tasks. Without proper regularization, autoencoder models are susceptible to the overfitting problem and the so-called posterior collapse phenomenon. In this paper, we introduce a quantization-based regularizer in the bottleneck stage of autoencoder models to learn meaningful latent representations. We combine both perspectives of Vector Quantized-Variational AutoEncoders (VQ-VAE) and classical denoising regularization methods of neural networks. We interpret quantizers as r…
University of Appl. Sciences Mittweida（米特韦达应用科学大学）；University Paris Sorbonne-Pantheon（巴黎索邦-先贤祠大学）
We propose in this contribution a method for l one regularization in prototype based relevance learning vector quantization (LVQ) for sparse relevance profiles. Sparse relevance profiles in hyperspectral data analysis fade down those spectral bands which are not necessary for classification. In particular, we consider the sparsity in the relevance profile enforced by LASSO optimization. The latter one is obtained by a gradient learning scheme using a differentiable parametrized approximation of the $l_{1}$-norm, which has an upper error bound. We extend this regularization idea also to the mat…
Robust quantization improves the tolerance of networks for various implementations, allowing reliable output in different bit-widths or fragmented low-precision arithmetic. In this work, we perform extensive analyses to identify the sources of quantization error and present three insights to robustify a network against quantization: reduction of error propagation, range clamping for error minimization, and inherited robustness against quantization. Based on these insights, we propose two novel methods called symmetry regularization (SymReg) and saturating nonlinearity (SatNL). Applying the pro…
We propose a simple and easy to implement neural network compression algorithm that achieves results competitive with more complicated state-of-the-art methods. The key idea is to modify the original optimization problem by adding K independent Gaussian priors (corresponding to the k-means objective) over the network parameters to achieve parameter quantization, as well as an L1 penalty to achieve pruning. Unlike many existing quantization-based methods, our method uses hard clustering assignments of network parameters, which adds minimal change or overhead to standard network training. We als…
Network quantization, which aims to reduce the bit-lengths of the network weights and activations, has emerged as one of the key ingredients to reduce the size of neural networks for their deployments to resource-limited devices. In order to overcome the nature of transforming continuous activations and weights to discrete ones, recent study called Relaxed Quantization (RQ) [Louizos et al. 2019] successfully employ the popular Gumbel-Softmax that allows this transformation with efficient gradient-based optimization. However, RQ with this Gumbel-Softmax relaxation still suffers from bias-varian…
Mitigating the Impact of Outlier Channels for Language Model Quantization with Activation Regularization
Massachusetts Institute of Technology（麻省理工学院）；MIT-IBM Watson AI Lab（MIT-IBM沃森人工智能实验室）；IBM Research（IBM研究院）；IST Austria（奥地利科学技术研究所）；Neural Magic, Inc.（Neural Magic公司）
For on-device automatic speech recognition (ASR), quantization aware training (QAT) is ubiquitous to achieve the trade-off between model predictive performance and efficiency. Among existing QAT methods, one major drawback is that the quantization centroids have to be predetermined and fixed. To overcome this limitation, we introduce a regularization-free, "soft-to-hard" compression mechanism with self-adjustable centroids in a mu-Law constrained space, resulting in a simpler yet more versatile quantization scheme, called General Quantizer (GQ). We apply GQ to ASR tasks using Recurrent Neural …
Astro: Activation-guided Structured Regularization for Outlier-Robust LLM Post-Training Quantization
Beijing Institute of Technology（北京理工大学）；Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；Hebei Province Key Laboratory of Big Data Science and Intelligent Technology（河北省大数据科学与智能技术重点实验室）
This paper describes an entropy regularization term for vector quantization (VQ) based on the analysis of persistent homology of the VQ embeddings. Higher embedding entropy positively correlates with higher codebook utilization, mitigating overfit towards the identity and codebook collapse in VQ-based autoencoders [1]. We show that homology-constrained regularization is an effective way to increase entropy of the VQ process (approximated to input entropy) while preserving the approximated topology in the quantized latent space, averaged over mini batches. This work further explores some patter…
Quantization of neural networks has become common practice, driven by the need for efficient implementations of deep neural networks on embedded devices. In this paper, we exploit an oft-overlooked degree of freedom in most networks - for a given layer, individual output channels can be scaled by any factor provided that the corresponding weights of the next layer are inversely scaled. Therefore, a given network has many factorizations which change the weights of the network without changing its function. We present a conceptually simple and easy to implement method that uses this property and…
We present a method for training a deep neural network containing sinusoidal activation functions to fit to time-series data. Weights are initialized using a fast Fourier transform, then trained with regularization to improve generalization. A simple dynamic parameter tuning method is employed to adjust both the learning rate and regularization term, such that stability and efficient training are both achieved. We show how deeper layers can be utilized to model the observed sequence using a sparser set of sinusoid units, and how non-uniform regularization can improve generalization by promotin…
Korea Advanced Institute of Science and Technology (KAIST)（韩国科学技术院）；AITRICS
Network quantization, which aims to reduce the bit-lengths of the network weights and activations, has emerged for their deployments to resource-limited devices. Although recent studies have successfully discretized a full-precision network, they still incur large quantization errors after training, thus giving rise to a significant performance gap between a full-precision network and its quantized counterpart. In this work, we propose a novel quantization method for neural networks, Cluster-Promoting Quantization (CPQ) that finds the optimal quantization grids while naturally encouraging the …
The Residual Quantization (RQ) framework is revisited where the quantization distortion is being successively reduced in multi-layers. Inspired by the reverse-water-filling paradigm in rate-distortion theory, an efficient regularization on the variances of the codewords is introduced which allows to extend the RQ for very large numbers of layers and also for high dimensional data, without getting over-trained. The proposed Regularized Residual Quantization (RRQ) results in multi-layer dictionaries which are additionally sparse, thanks to the soft-thresholding nature of the regularization when …
Neural network models are resource hungry. It is difficult to deploy such deep networks on devices with limited resources, like smart wearables, cellphones, drones, and autonomous vehicles. Low bit quantization such as binary and ternary quantization is a common approach to alleviate this resource requirements. Ternary quantization provides a more flexible model and outperforms binary quantization in terms of accuracy, however doubles the memory footprint and increases the computational cost. Contrary to these approaches, mixed quantized models allow a trade-off between accuracy and memory foo…
Huawei Noah’s Ark Lab（华为诺亚方舟实验室）；University of Montreal（蒙特利尔大学）
Deep neural networks (DNNs) have demonstrated success for many supervised learning tasks, ranging from voice recognition, object detection, to image classification. However, their increasing complexity might yield poor generalization error that make them hard to be deployed on edge devices. Quantization is an effective approach to compress DNNs in order to meet these constraints. Using a quasiconvex base function in order to construct a binary quantizer helps training binary neural networks (BNNs) and adding noise to the input data or using a concrete regularization function helps to improve g…
Deep learning models have been successfully used in computer vision and many other fields. We propose an unorthodox algorithm for performing quantization of the model parameters. In contrast with popular quantization schemes based on thresholds, we use a novel technique based on periodic functions, such as continuous trigonometric sine or cosine as well as non-continuous hat functions. We apply these functions component-wise and add the sum over the model parameters as a regularizer to the model loss during training. The frequency and amplitude hyper-parameters of these functions can be adjust…
We consider learning deep neural networks (DNNs) that consist of low-precision weights and activations for efficient inference of fixed-point operations. In training low-precision networks, gradient descent in the backward pass is performed with high-precision weights while quantized low-precision weights and activations are used in the forward pass to calculate the loss function for training. Thus, the gradient descent becomes suboptimal, and accuracy loss follows. In order to reduce the mismatch in the forward and backward passes, we utilize mean squared quantization error (MSQE) regularizat…
We propose a Quantization Guided Training (QGT) method to guide DNN training towards optimized low-bit-precision targets and reach extreme compression levels below 8-bit precision. Unlike standard quantization-aware training (QAT) approaches, QGT uses customized regularization to encourage weight values towards a distribution that maximizes accuracy while reducing quantization errors. One of the main benefits of this approach is the ability to identify compression bottlenecks. We validate QGT using state-of-the-art model architectures on vision datasets. We also demonstrate the effectiveness o…
Weight quantization is one of the most important techniques of Deep Neural Networks (DNNs) model compression method. A recent work using systematic framework of DNN weight quantization with the advanced optimization algorithm ADMM (Alternating Direction Methods of Multipliers) achieves one of state-of-art results in weight quantization. In this work, we first extend such ADMM-based framework to guarantee solution feasibility and we have further developed a multi-step, progressive DNN weight quantization framework, with dual benefits of (i) achieving further weight quantization thanks to the sp…
Universitat de Barcelona（巴塞罗那大学）；Bucknell University（巴克内尔大学）
Neural networks are more expressive when they have multiple layers. In turn, conventional training methods are only successful if the depth does not lead to numerical issues such as exploding or vanishing gradients, which occur less frequently when the layers are sufficiently wide. However, increasing width to attain greater depth entails the use of heavier computational resources and leads to overparameterized models. These subsequent issues have been partially addressed by model compression methods such as quantization and pruning, some of which relying on normalization-based regularization …
University of Toronto（多伦多大学）；York University（约克大学）
In this work, PRUNIX, a framework for training and pruning convolutional neural networks is proposed for deployment on memristor crossbar based accelerators. PRUNIX takes into account the numerous non-ideal effects of memristor crossbars including weight quantization, state-drift, aging and stuck-at-faults. PRUNIX utilises a novel Group Sawtooth Regularization intended to improve non-ideality tolerance as well as sparsity, and a novel Adaptive Pruning Algorithm (APA) intended to minimise accuracy loss by considering the sensitivity of different layers of a CNN to pruning. We compare our regula…
Seoul National University（首尔国立大学）；Qualcomm AI Research（高通人工智能研究院）；Qualcomm Technologies, Inc.（高通技术公司）
Quantization and Knowledge distillation (KD) methods are widely used to reduce memory and power consumption of deep neural networks (DNNs), especially for resource-constrained edge devices. Although their combination is quite promising to meet these requirements, it may not work as desired. It is mainly because the regularization effect of KD further diminishes the already reduced representation power of a quantized model. To address this short-coming, we propose Quantization-aware Knowledge Distillation (QKD) wherein quantization and KD are care-fully coordinated in three phases. First, Self-…
Fermi National Accelerator Laboratory（费米国家加速器实验室）；University of California San Diego（加州大学圣地亚哥分校）；Xilinx Research（赛灵思研究院）；Northwestern University（西北大学）
Efficient machine learning implementations optimized for inference in hardware have wide-ranging benefits, depending on the application, from lower inference latency to higher data throughput and reduced energy consumption. Two popular techniques for reducing computation in neural networks are pruning, removing insignificant synapses, and quantization, reducing the precision of the calculations. In this work, we explore the interplay between pruning and quantization during the training of neural networks for ultra low latency applications targeting high energy physics use cases. Techniques dev…
