Efficient similarity retrieval from large-scale multimodal database is pervasive in modern search engines and social networks. To support queries across content modalities, the system should enable cross-modal correlation and computation-efficient indexing. While hashing methods have shown great potential in achieving this goal, current attempts generally fail to learn isomorphic hash codes in a seamless scheme, that is, they embed multiple modalities in a continuous isomorphic space and separately threshold embeddings into binary codes, which incurs substantial loss of retrieval accuracy. In this paper, we approach seamless multimodal hashing by proposing a novel Composite Correlation Quantization (CCQ) model. Specifically, CCQ jointly finds correlation-maximal mappings that transform different modalities into isomorphic latent space, and learns composite quantizers that convert the isomorphic latent features into compact binary codes. An optimization framework is devised to preserve both intra-modal similarity and inter-modal correlation through minimizing both reconstruction and quantization errors, which can be trained from both paired and partially paired data in linear time. A comprehensive set of experiments clearly show the superior effectiveness and efficiency of CCQ against the state of the art hashing methods for both unimodal and cross-modal retrieval.
A central aspect of online decision tree solutions is evaluating the incoming data and enabling model growth. For such, trees much deal with different kinds of input features and partition them to learn from the data. Numerical features are no exception, and they pose additional challenges compared to other kinds of features, as there is no trivial strategy to choose the best point to make a split decision. The problem is even more challenging in regression tasks because both the features and the target are continuous. Typical online solutions evaluate and store all the points monitored between split attempts, which goes against the constraints posed in real-time applications. In this paper, we introduce the Quantization Observer (QO), a simple yet effective hashing-based algorithm to monitor and evaluate split point candidates in numerical features for online tree regressors. QO can be easily integrated into incremental decision trees, such as Hoeffding Trees, and it has a monitoring cost of $O(1)$ per instance and sub-linear cost to evaluate split candidates. Previous solutions had a $O(\log n)$ cost per insertion (in the best case) and a linear cost to evaluate split points. Our extensive experimental setup highlights QO's effectiveness in providing accurate split point suggestions while spending much less memory and processing time than its competitors.
Estimating the primary quantization matrix of double JPEG compressed images is a problem of relevant importance in image forensics since it allows to infer important information about the past history of an image. In addition, the inconsistencies of the primary quantization matrices across different image regions can be used to localize splicing in double JPEG tampered images. Traditional model-based approaches work under specific assumptions on the relationship between the first and second compression qualities and on the alignment of the JPEG grid. Recently, a deep learning-based estimator capable to work under a wide variety of conditions has been proposed, that outperforms tailored existing methods in most of the cases. The method is based on a Convolutional Neural Network (CNN) that is trained to solve the estimation as a standard regression problem. By exploiting the integer nature of the quantization coefficients, in this paper, we propose a deep learning technique that performs the estimation by resorting to a simil-classification architecture. The CNN is trained with a loss function that takes into account both the accuracy and the Mean Square Error (MSE) of the estimation. Results confirm the superior performance of the proposed technique, compared to the state-of-the art methods based on statistical analysis and, in particular, deep learning regression. Moreover, the capability of the method to work under general operative conditions, regarding the alignment of the second compression grid with the one of first compression and the combinations of the JPEG qualities of former and second compression, is very relevant in practical applications, where these information are unknown a priori.
Communication constraints are one of the major challenges preventing the wide-spread adoption of Federated Learning systems. Recently, Federated Distillation (FD), a new algorithmic paradigm for Federated Learning with fundamentally different communication properties, emerged. FD methods leverage ensemble distillation techniques and exchange model outputs, presented as soft labels on an unlabeled public data set, between the central server and the participating clients. While for conventional Federated Learning algorithms, like Federated Averaging (FA), communication scales with the size of the jointly trained model, in FD communication scales with the distillation data set size, resulting in advantageous communication properties, especially when large models are trained. In this work, we investigate FD from the perspective of communication efficiency by analyzing the effects of active distillation-data curation, soft-label quantization and delta-coding techniques. Based on the insights gathered from this analysis, we present Compressed Federated Distillation (CFD), an efficient Federated Distillation method. Extensive experiments on Federated image classification and language modeling problems demonstrate that our method can reduce the amount of communication necessary to achieve fixed performance targets by more than two orders of magnitude, when compared to FD and by more than four orders of magnitude when compared with FA.
Edge computing is promising to become one of the next hottest topics in artificial intelligence because it benefits various evolving domains such as real-time unmanned aerial systems, industrial applications, and the demand for privacy protection. This paper reviews recent advances on binary neural network (BNN) and 1-bit CNN technologies that are well suitable for front-end, edge-based computing. We introduce and summarize existing work and classify them based on gradient approximation, quantization, architecture, loss functions, optimization method, and binary neural architecture search. We also introduce applications in the areas of computer vision and speech recognition and discuss future applications for edge computing.
With the rising popularity of intelligent mobile devices, it is of great practical significance to develop accurate, realtime and energy-efficient image Super-Resolution (SR) inference methods. A prevailing method for improving the inference efficiency is model quantization, which allows for replacing the expensive floating-point operations with efficient fixed-point or bitwise arithmetic. To date, it is still challenging for quantized SR frameworks to deliver feasible accuracy-efficiency trade-off. Here, we propose a Fully Quantized image Super-Resolution framework (FQSR) to jointly optimize efficiency and accuracy. In particular, we target on obtaining end-to-end quantized models for all layers, especially including skip connections, which was rarely addressed in the literature. We further identify training obstacles faced by low-bit SR networks and propose two novel methods accordingly. The two difficulites are caused by 1) activation and weight distributions being vastly distinctive in different layers; 2) the inaccurate approximation of the quantization. We apply our quantization scheme on multiple mainstream super-resolution architectures, including SRResNet, SRGAN and EDSR. Experimental results show that our FQSR using low bits quantization can achieve on par performance compared with the full-precision counterparts on five benchmark datasets and surpass state-of-the-art quantized SR methods with significantly reduced computational cost and memory consumption.
Transformer-based language models such as BERT provide significant accuracy improvement for a multitude of natural language processing (NLP) tasks. However, their hefty computational and memory demands make them challenging to deploy to resource-constrained edge platforms with strict latency requirements. We present EdgeBERT, an in-depth algorithm-hardware co-design for latency-aware energy optimization for multi-task NLP. EdgeBERT employs entropy-based early exit predication in order to perform dynamic voltage-frequency scaling (DVFS), at a sentence granularity, for minimal energy consumption while adhering to a prescribed target latency. Computation and memory footprint overheads are further alleviated by employing a calibrated combination of adaptive attention span, selective network pruning, and floating-point quantization. Furthermore, in order to maximize the synergistic benefits of these algorithms in always-on and intermediate edge computing settings, we specialize a 12nm scalable hardware accelerator system, integrating a fast-switching low-dropout voltage regulator (LDO), an all-digital phase-locked loop (ADPLL), as well as, high-density embedded non-volatile memories (eNVMs) wherein the sparse floating-point bit encodings of the shared multi-task parameters are carefully stored. Altogether, latency-aware multi-task NLP inference acceleration on the EdgeBERT hardware system generates up to 7x, 2.5x, and 53x lower energy compared to the conventional inference without early stopping, the latency-unbounded early exit approach, and CUDA adaptations on an Nvidia Jetson Tegra X2 mobile GPU, respectively.
In recent years, the growing size of neural networks has led to a vast amount of research concerning compression techniques to mitigate the drawbacks of such large sizes. Most of these research works can be categorized into three broad families : Knowledge Distillation, Pruning, and Quantization. While there has been steady research in this domain, adoption and commercial usage of the proposed techniques has not quite progressed at the rate. We present KD-Lib, an open-source PyTorch based library, which contains state-of-the-art modular implementations of algorithms from the three families on top of multiple abstraction layers. KD-Lib is model and algorithm-agnostic, with extended support for hyperparameter tuning using Optuna and Tensorboard for logging and monitoring. The library can be found at - https://github.com/SforAiDl/KD_Lib.
Depth factorization and quantization have emerged as two of the principal strategies for designing efficient deep convolutional neural network (CNN) architectures tailored for low-power inference on the edge. However, there is still little detailed understanding of how different depth factorization choices affect the final, trained distributions of each layer in a CNN, particularly in the situation of quantized weights and activations. In this study, we introduce a progressive depth factorization strategy for efficient CNN architecture exploration under quantization constraints. By algorithmically increasing the granularity of depth factorization in a progressive manner, the proposed strategy enables a fine-grained, low-level analysis of layer-wise distributions. Thus enabling the gain of in-depth, layer-level insights on efficiency-accuracy tradeoffs under fixed-precision quantization. Such a progressive depth factorization strategy also enables efficient identification of the optimal depth-factorized macroarchitecture design (which we will refer to here as FactorizeNet) based on the desired efficiency-accuracy requirements.
With the proliferation of deep convolutional neural network (CNN) algorithms for mobile processing, limited precision quantization has become an essential tool for CNN efficiency. Consequently, various works have sought to design fixed precision quantization algorithms and quantization-focused optimization techniques that minimize quantization induced performance degradation. However, there is little concrete understanding of how various CNN design decisions/best practices affect quantized inference behaviour. Weight initialization strategies are often associated with solving issues such as vanishing/exploding gradients but an often-overlooked aspect is their impact on the final trained distributions of each layer. We present an in-depth, fine-grained ablation study of the effect of different weights initializations on the final distributions of weights and activations of different CNN architectures. The fine-grained, layerwise analysis enables us to gain deep insights on how initial weights distributions will affect final accuracy and quantized behaviour. To our best knowledge, we are the first to perform such a low-level, in-depth quantitative analysis of weights initialization and its effect on quantized behaviour.
Semantic segmentation has been a major topic in research and industry in recent years. However, due to the computation complexity of pixel-wise prediction and backpropagation algorithm, semantic segmentation has been demanding in computation resources, resulting in slow training and inference speed and large storage space to store models. Existing schemes that speed up segmentation network change the network structure and come with noticeable accuracy degradation. However, neural network quantization can be used to reduce computation load while maintaining comparable accuracy and original network structure. Semantic segmentation networks are different from traditional deep convolutional neural networks (DCNNs) in many ways, and this topic has not been thoroughly explored in existing works. In this paper, we propose a new quantization framework for training and inference of segmentation networks, where parameters and operations are constrained to 8-bit integer-based values for the first time. Full quantization of the data flow and the removal of square and root operations in batch normalization give our framework the ability to perform inference on fixed-point devices. Our proposed framework is evaluated on mainstream semantic segmentation networks like FCN-VGG16 and DeepLabv3-ResNet50, achieving comparable accuracy against floating-point framework on ADE20K dataset and PASCAL VOC 2012 dataset.
The world is going through one of the most dangerous pandemics of all time with the rapid spread of the novel coronavirus (COVID-19). According to the World Health Organisation, the most effective way to thwart the transmission of coronavirus is to wear medical face masks. Monitoring the use of face masks in public places has been a challenge because manual monitoring could be unsafe. This paper proposes an architecture for detecting medical face masks for deployment on resource-constrained endpoints having extremely low memory footprints. A small development board with an ARM Cortex-M7 microcontroller clocked at 480 Mhz and having just 496 KB of framebuffer RAM, has been used for the deployment of the model. Using the TensorFlow Lite framework, the model is quantized to further reduce its size. The proposed model is 138 KB post quantization and runs at the inference speed of 30 FPS.
Due to the boom in technical compute in the last few years, the world has seen massive advances in artificially intelligent systems solving diverse real-world problems. But a major roadblock in the ubiquitous acceptance of these models is their enormous computational complexity and memory footprint. Hence efficient architectures and training techniques are required for deployment on extremely low resource inference endpoints. This paper proposes an architecture for detection of alphabets in American Sign Language on an ARM Cortex-M7 microcontroller having just 496 KB of framebuffer RAM. Leveraging parameter quantization is a common technique that might cause varying drops in test accuracy. This paper proposes using interpolation as augmentation amongst other techniques as an efficient method of reducing this drop, which also helps the model generalize well to previously unseen noisy data. The proposed model is about 185 KB post-quantization and inference speed is 20 frames per second.
While deep learning has achieved phenomenal successes in many AI applications, its enormous model size and intensive computation requirements pose a formidable challenge to the deployment in resource-limited nodes. There has recently been an increasing interest in computationally-efficient learning methods, e.g., quantization, pruning and channel gating. However, most existing techniques cannot adapt to different tasks quickly. In this work, we advocate a holistic approach to jointly train the backbone network and the channel gating which enables dynamical selection of a subset of filters for more efficient local computation given the data input. Particularly, we develop a federated meta-learning approach to jointly learn good meta-initializations for both backbone networks and gating modules, by making use of the model similarity across learning tasks on different nodes. In this way, the learnt meta-gating module effectively captures the important filters of a good meta-backbone network, based on which a task-specific conditional channel gated network can be quickly adapted, i.e., through one-step gradient descent, from the meta-initializations in a two-stage procedure using new samples of that task. The convergence of the proposed federated meta-learning algorithm is established under mild conditions. Experimental results corroborate the effectiveness of our method in comparison to related work.
Radar sensors are crucial for environment perception of driver assistance systems as well as autonomous vehicles. Key performance factors are weather resistance and the possibility to directly measure velocity. With a rising number of radar sensors and the so far unregulated automotive radar frequency band, mutual interference is inevitable and must be dealt with. Algorithms and models operating on radar data in early processing stages are required to run directly on specialized hardware, i.e. the radar sensor. This specialized hardware typically has strict resource-constraints, i.e. a low memory capacity and low computational power. Convolutional Neural Network (CNN)-based approaches for denoising and interference mitigation yield promising results for radar processing in terms of performance. However, these models typically contain millions of parameters, stored in hundreds of megabytes of memory, and require additional memory during execution. In this paper we investigate quantization techniques for CNN-based denoising and interference mitigation of radar signals. We analyze the quantization potential of different CNN-based model architectures and sizes by considering (i) quantized weights and (ii) piecewise constant activation functions, which results in reduced memory requirements for model storage and during the inference step respectively.
In rate-distortion optimization, the encoder settings are determined by maximizing a reconstruction quality measure subject to a constraint on the bit rate. One of the main challenges of this approach is to define a quality measure that can be computed with low computational cost and which correlates well with perceptual quality. While several quality measures that fulfil these two criteria have been developed for images and video, no such one exists for 3D point clouds. We address this limitation for the video-based point cloud compression (V-PCC) standard by proposing a linear perceptual quality model whose variables are the V-PCC geometry and color quantization parameters and whose coefficients can easily be computed from two features extracted from the original 3D point cloud. Subjective quality tests with 400 compressed 3D point clouds show that the proposed model correlates well with the mean opinion score, outperforming state-of-the-art full reference objective measures in terms of Spearman rank-order and Pearsons linear correlation coefficient. Moreover, we show that for the same target bit rate, ratedistortion optimization based on the proposed model offers higher perceptual quality than rate-distortion optimization based on exhaustive search with a point-to-point objective quality metric.
Visual quality evaluation is one of the challenging basic problems in image processing. It also plays a central role in the shaping, implementation, optimization, and testing of many methods. The existing image quality assessment methods focused on images corrupted by common degradation types while little attention was paid to color quantization. This in spite there is a wide range of applications requiring color quantization assessment being used as a preprocessing step when color-based tasks are more efficiently accomplished on a reduced number of colors. In this paper, we propose and carry-out a quantitative performance evaluation of nine well-known and commonly used full-reference image quality assessment measures. The evaluation is done by using two publicly available and subjectively rated image quality databases for color quantization degradation and by considering suitable combinations or subparts of them. The results indicate the quality measures that have closer performances in terms of their correlation to the subjective human rating and show that the evaluation of the statistical performance of the quality measures for color quantization is significantly impacted by the selected image quality database while maintaining a similar trend on each database. The detected strong similarity both on individual databases and on databases obtained by integration provides the ability to validate the integration process and to consider the quantitative performance evaluation on each database as an indicator for performance on the other databases. The experimental results are useful to address the choice of suitable quality measures for color quantization and to improve their future employment.
Current low-precision quantization algorithms often have the hidden cost of conversion back and forth from floating point to quantized integer values. This hidden cost limits the latency improvement realized by quantizing Neural Networks. To address this, we present HAWQV3, a novel mixed-precision integer-only quantization framework. The contributions of HAWQV3 are the following: (i) An integer-only inference where the entire computational graph is performed only with integer multiplication, addition, and bit shifting, without any floating point operations or even integer division; (ii) A novel hardware-aware mixed-precision quantization method where the bit-precision is calculated by solving an integer linear programming problem that balances the trade-off between model perturbation and other constraints, e.g., memory footprint and latency; (iii) Direct hardware deployment and open source contribution for 4-bit uniform/mixed-precision quantization in TVM, achieving an average speed up of $1.45\times$ for uniform 4-bit, as compared to uniform 8-bit for ResNet50 on T4 GPUs; and (iv) extensive evaluation of the proposed methods on ResNet18/50 and InceptionV3, for various model compression levels with/without mixed precision. For ResNet50, our INT8 quantization achieves an accuracy of $77.58\%$, which is $2.68\%$ higher than prior integer-only work, and our mixed-precision INT4/8 quantization can reduce INT8 latency by $23\%$ and still achieve $76.73\%$ accuracy. Our framework and the TVM implementation have been open sourced.
WaveNet is a state-of-the-art text-to-speech vocoder that remains challenging to deploy due to its autoregressive loop. In this work we focus on ways to accelerate the original WaveNet architecture directly, as opposed to modifying the architecture, such that the model can be deployed as part of a scalable text-to-speech system. We survey a wide variety of model compression techniques that are amenable to deployment on a range of hardware platforms. In particular, we compare different model sparsity methods and levels, and seven widely used precisions as targets for quantization; and are able to achieve models with a compression ratio of up to 13.84 without loss in audio fidelity compared to a dense, single-precision floating-point baseline. All techniques are implemented using existing open source deep learning frameworks and libraries to encourage their wider adoption.
User data confidentiality protection is becoming a rising challenge in the present deep learning research. Without access to data, conventional data-driven model compression faces a higher risk of performance degradation. Recently, some works propose to generate images from a specific pretrained model to serve as training data. However, the inversion process only utilizes biased feature statistics stored in one model and is from low-dimension to high-dimension. As a consequence, it inevitably encounters the difficulties of generalizability and inexact inversion, which leads to unsatisfactory performance. To address these problems, we propose MixMix based on two simple yet effective techniques: (1) Feature Mixing: utilizes various models to construct a universal feature space for generalized inversion; (2) Data Mixing: mixes the synthesized images and labels to generate exact label information. We prove the effectiveness of MixMix from both theoretical and empirical perspectives. Extensive experiments show that MixMix outperforms existing methods on the mainstream compression tasks, including quantization, knowledge distillation, and pruning. Specifically, MixMix achieves up to 4% and 20% accuracy uplift on quantization and pruning, respectively, compared to existing data-free compression work.