EC-DNN: A new method for parallel training of deep neural networks

Federated learning (FL) enables decentralized training of deep neural networks (DNNs) for object detection over a distributed population of clients. It allows edge clients to keep their data local and only share parameter updates with a federated server. However, the distributed nature of FL also opens doors to new threats. In this paper, we present targeted perception poisoning attacks against federated object detection learning in which a subset of malicious clients seeks to poison the federated training of a global object detection model by sharing perception-poisoned local model parameters. We first introduce three targeted perception poisoning attacks, which have severe adverse effects only on the objects under attack. We then analyze the attack feasibility, the impact of malicious client availability, and attack timing. To safeguard FL systems against such contagious threats, we introduce spatial signature analysis as a defense to separate benign local model parameters from poisoned local model contributions, identify malicious clients, and eliminate their impact on the federated training. Extensive experiments on object detection benchmark datasets validate that the defense-empowered federated object detection learning can improve the robustness against all three types of perception poisoning attacks. The source code is available at https://github.com/git-disl/Perception-Poisoning.

Deep Learning (DL) has become a prominent machine learning technique due to the availability of efficient computational resources in the form of Graphics Processing Units (GPUs), large-scale datasets and a variety of models. The newer generation of GPUs are being designed with special emphasis on optimizing performance for DL applications. Also, the availability of easy-to-use DL frameworks—like PyTorch and TensorFlow— has enhanced productivity of domain experts to work on their custom DL applications from diverse domains. However, existing Deep Neural Network (DNN) training approaches may not fully utilize the newly emerging powerful GPUs like the NVIDIA A100—this is the primary issue that we address in this paper. Our motivating analyses show that the GPU utilization on NVIDIA A100 can be as low as 43% using traditional DNN training approaches for small-to-medium DL models and input data size. This paper proposes AccDP—a data-parallel distributed DNN training approach—to accelerate GPU-based DL applications. AccDP exploits the Message Passing Interface (MPI) communication library coupled with the NVIDIA’s Multi-Process Service (MPS) to increase the amount of work assigned to parallel GPUs resulting in higher utilization of compute resources. We evaluate our proposed design on different small-to-medium DL models and input sizes on the state-of-the-art HPC clusters. By injecting more parallelism into DNN training using our approach, the evaluation shows up to 58% improvement in training performance on a single GPU and up to 62% on 16 GPUs compared to regular DNN training. Furthermore, we conduct an in-depth characterization to determine the impact of several DNN training factors and best practices—including the batch size and the number of data loading workers— to optimally utilize GPU devices. To the best of our knowledge, this is the first work that explores the use of MPS and MPI to maximize the utilization of GPUs in distributed DNN training.

Recently deep-learning (DL) techniques have been widely adopted in side-channel power analysis. A DL-assisted SCA generally consists of two phases: a deep neural network (DNN) training phase and a follow-on attack phase using the trained DNN. However, currently the two phases are not well aligned, as there is no conclusion on what metric used in the training can result in the most effective attack in the second phase. When traditional loss functions such as negative log-likelihood (NLL) are used in training a DNN, the trained model does not yield optimal follow-on attack. Recently some information theoretical SCA leakage metrics are proposed, either as the validation metric to stop the DNN training with traditional loss functions, or as both the validation metric and the training loss function. None of those proposed metrics, however, directly measures the SCA effectiveness. We propose to conduct DNN training directly with a common SCA effectiveness metric, Guessing Entropy (GE). We overcome the prior practical difficulty of using GE in DNN training by utilizing the GEEA estimation algorithm introduced in CHES 2020. We show that using GEEA as either the validation metric or the loss function produces DNN models that lead to much more effective follow-on attacks. Our work consolidates the DL-assisted SCA framework with a consistent metric, which shows great potential to be adopted as the universal SCA-oriented DNN training framework.

In this invited paper, we present deep neural network (DNN) training accelerator designs in both ASIC and FPGA. The accelerators implements stochastic gradient descent based training algorithm in 16-bit fixed-point precision. A new cyclic weight storage and access scheme enables using the same off-the-shelf SRAMs for non-transpose and transpose operations during feed-forward and feed-backward phases, respectively, of the DNN training process. Including the cyclic weight scheme, the overall DNN training processor is implemented in both 65nm CMOS ASIC and Intel Stratix-10 FPGA hardware. We collectively report the ASIC and FPGA training accelerator results.

In federated learning (FL), the communication constraint between the remote clients and the Parameter Server (PS) is a crucial bottleneck. For this reason, model updates must be compressed so as to minimize the loss in accuracy resulting from the communication constraint. This paper proposes “M-magnitude weighted L2 distortion + 2 degrees of freedom” (M22) algorithm, a rate-distortion inspired approach to gradient compression for federated training of deep neural networks (DNNs). In particular, we propose a family of distortion measures between the original gradient and the reconstruction we referred to as “ <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$M$ </tex-math></inline-formula> -magnitude weighted <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$L_{2}$ </tex-math></inline-formula> ” distortion, and we assume that gradient updates follow an i.i.d. distribution – generalized normal or Weibull, which have two degrees of freedom. In both the distortion measure and the gradient distribution, there is one free parameter for each that can be fitted as a function of the iteration number. Given a choice of gradient distribution and distortion measure, we design the quantizer to minimize the expected distortion in gradient reconstruction. To measure the gradient compression performance under a communication constraint, we define the per-bit accuracy as the optimal improvement in accuracy that one bit of communication brings to the centralized model over the training period. Using this performance measure, we systematically benchmark the choice of gradient distribution and distortion measure. We provide substantial insights on the role of these choices and argue that significant performance improvements can be attained using such a rate-distortion inspired compressor.

Survey on Deep Neural Networks in Speech and Vision Systems

Previous work presented a proof of concept for sequence training of deep neural networks (DNNs) using asynchronous stochastic optimization, mainly focusing on a small-scale task. The approach offers the potential to leverage both the efficiency of stochastic gradient descent and the scalability of parallel computation. This study presents results for four different voice search tasks to confirm the effectiveness and efficiency of the proposed framework across different conditions: amount of data (from 60 hours to 20,000 hours), type of speech (read speech vs. spontaneous speech), quality of data (supervised vs. unsupervised data), and language. Significant gains over baselines (DNNs trained at the frame level) are found to hold across these conditions. The experimental results are analyzed, and additional practical details for the approach are provided. Furthermore, different sequence training criteria are compared.

The performance and efficiency of distributed training of Deep Neural Networks (DNN) highly depend on the performance of gradient averaging among participating processes, a step bound by communication costs. There are two major approaches to reduce communication overhead: overlap communications with computations (lossless), or reduce communications (lossy). The lossless solution works well for linear neural architectures, e.g. VGG, AlexNet, but more recent networks such as ResNet and Inception limit the opportunity for such overlapping. Therefore, approaches that reduce the amount of data (lossy) become more suitable. In this paper, we present a novel, explainable lossy method that sparsifies gradients in the frequency domain, in addition to a new range-based float point representation to quantize and further compress gradients. These dynamic techniques strike a balance between compression ratio, accuracy, and computational overhead, and are optimized to maximize performance in heterogeneous environments.

As the popularity of deep learning in industry rapidly grows, efficient training of deep neural networks (DNNs) becomes important. To train a DNN with a large amount of data, distributed training with data parallelism has been widely adopted. However, the communication overhead limits the scalability of distributed training. To reduce the overhead, a number of distributed training algorithms have been proposed. The model accuracy and training performance of those algorithms can be different depending on various factors such as cluster settings, training models/datasets, and optimization techniques applied. In order for someone to adopt a distributed training algorithm appropriate for her/his situation, it is required for her/him to fully understand the model accuracy and training performance of these algorithms in various settings. Toward this end, this paper reviews and evaluates seven popular distributed training algorithms (BSP, ASP, SSP, EASGD, AR-SGD, GoSGD, and AD-PSGD) in terms of the model accuracy and training performance in various settings. Specifically, we evaluate those algorithms for two CNN models, in different cluster settings, and with three well-known optimization techniques. Through extensive evaluation and analysis, we made several interesting discoveries. For example, we found out that some distributed training algorithms (SSP, EASGD, and GoSGD) have highly negative impact on the model accuracy because they adopt intermittent and asymmetric communication to improve training performance; the communication overhead of some centralized algorithms (ASP and SSP) is much higher than we expected in a cluster setting with limited network bandwidth because of the PS bottleneck problem. These findings, and many more in the paper, can guide the adoption of proper distributed training algorithms in industry; our findings can be useful in academia as well for designing new distributed training algorithms.

Contact-map of a protein sequence dictates the global topology of structural fold. Accurate prediction of the contact-map is thus essential to protein 3D structure prediction, which is particularly useful for the protein sequences that do not have close homology templates in the Protein Data Bank. We developed a new method, ResPRE, to predict residue-level protein contacts using inverse covariance matrix (or precision matrix) of multiple sequence alignments (MSAs) through deep residual convolutional neural network training. The approach was tested on a set of 158 non-homologous proteins collected from the CASP experiments and achieved an average accuracy of 50.6% in the top-L long-range contact prediction with L being the sequence length, which is 11.7% higher than the best of other state-of-the-art approaches ranging from coevolution coupling analysis to deep neural network training. Detailed data analyses show that the major advantage of ResPRE lies at the utilization of precision matrix that helps rule out transitional noises of contact-maps compared with the previously used covariance matrix. Meanwhile, the residual network with parallel shortcut layer connections increases the learning ability of deep neural network training. It was also found that appropriate collection of MSAs can further improve the accuracy of final contact-map predictions. The standalone package and online server of ResPRE are made freely available, which should bring important impact on protein structure and function modeling studies in particular for the distant- and non-homology protein targets. https://zhanglab.ccmb.med.umich.edu/ResPRE and https://github.com/leeyang/ResPRE. Supplementary data are available at Bioinformatics online.

&lt;p style="text-indent: 1.27cm; margin-bottom: 0.35cm; line-height: 115%;" align="justify"&gt;&lt;span style="font-family: Arial,serif;"&gt;&lt;span style="font-size: small;"&gt;&lt;em&gt;Deep Neural Network training algorithms consumes long training time, especially when the number of hidden layers and nodes is large. Matrix multiplication is the key operation carried out at every node of each layer for several hundreds of thousands of times during the training of Deep Neural Network. Blocking is a well-proven optimization technique to improve the performance of matrix multiplication. Blocked Matrix multiplication algorithms can easily be parallelized to accelerate the performance further. This paper proposes a novel approach of implementing Parallel Blocked Matrix multiplication algorithms to reduce the long training time. The proposed approach was implemented using a parallel programming model OpenMP with collapse() clause for the multiplication of input and weight matrices of Backpropagation and Boltzmann Machine Algorithms for training Deep Neural Network and tested on multi-core processor system. Experimental results showed that the proposed approach achieved approximately two times speedup than classic algorithms.&lt;/em&gt;&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;

Graphics processing units (GPUs) have been widely adopted to accelerate the training of deep neural networks (DNNs). Although the computational performance of GPUs has been improving steadily, the memory size of modern GPUs is still quite limited, which restricts the sizes of DNNs that can be trained on GPUs, and hence raises serious challenges. This paper introduces a framework, referred to as moDNN (memory optimal DNN training on GPUs), to optimize the memory usage in DNN training. moDNN supports automatic tuning of DNN training code to match any given memory budget (not smaller than the theoretical lower bound). By taking full advantage of overlapping computations and data transfers, we develop new heuristics to judiciously schedule data offloading and prefetching transfers, together with convolution algorithm selection, to optimize memory usage. We further devise a new sub-batch size selection method which also greatly reduces memory usage. moDNN can save memory usage up to 59×, compared with an ideal case which assumes that the GPU memory is sufficient to hold all data. When executing moDNN on a GPU with 12 GB memory, the training time is increased by only 3 percent, which is much shorter than that incurred by the best known approach, vDNN. Furthermore, we propose an optimization strategy for moDNN on multiple GPUs again by utilizing the idea of overlapping data transfers and GPU computations. The results show that 3.7× speedup is attained on four GPUs.

Over the last decade, the deep neural networks are a hot topic in machine learning. It is breakthrough technology in processing images, video, speech, text and audio. Deep neural network permits us to overcome some limitations of a shallow neural network due to its deep architecture. In this paper we investigate the nature of unsupervised learning in restricted Boltzmann machine. We have proved that maximization of the log-likelihood input data distribution of restricted Boltzmann machine is equivalent to minimizing the cross-entropy and to special case of minimizing the mean squared error. Thus the nature of unsupervised learning is invariant to different training criteria. As a result we propose a new technique called "REBA" for the unsupervised training of deep neural networks. In contrast to Hinton's conventional approach to the learning of restricted Boltzmann machine, which is based on linear nature of training rule, the proposed technique is founded on nonlinear training rule. We have shown that the classical equations for RBM learning are a special case of the proposed technique. As a result the proposed approach is more universal in contrast to the traditional energy-based model. We demonstrate the performance of the REBA technique using wellknown benchmark problem. The main contribution of this paper is a novel view and new understanding of an unsupervised learning in deep neural networks.

The prediction of stock index movement is considered a rather significant objective in the financial world, since a reasonably accurate prediction has the possibility of gaining profit in stock exchange, yielding high financial benefits and hedging against market risks. Undoubtedly, the area of financial analysis has been dramatically changed from a rather qualitative science to a more quantitative science which is also based on knowledge extraction from databases. During the last years, deep learning constitutes a significant prediction tool in analyzing and exploiting the knowledge acquired from financial data. In this paper, we propose a new Deep Neural Network (DNN) prediction model for forecasting stock exchange index movement. The proposed DNN is characterized by the application of conditions on the weights in the form of box-constraints, during the training process. The motivation for placing these constraints is focused on defining the weights in the trained network in more uniform way, by restricting them from taking large values in order for all inputs and neurons of the DNN to be efficiently exploited and explored. The training of the new DNN model is performed by a Weight-Constrained Deep Neural Network (WCDNN) training algorithm which exploits the numerical efficiency and very low memory requirements of the L-BFGS (Limited-memory Broyden-Fletcher-Goldfarb-Shanno) matrices together with a gradient-projection strategy for handling the bounds on the weights of the network. The performance evaluation carried out on three popular stock exchange indices, demonstrates the classification efficiency of the proposed algorithm.

Only weights and biases are learned by gradient descent-based training of Deep Neural Networks (DNN). The other parameters, i.e., hyperparameters have a huge influence on the quality of the model but finding optimal values for them is not a trivial solution. The hyperparameter space grows exponentially and they also display non-linearity and interactions. A network with twelve hyperparameters, each with five potential values can have about 250 million unique combinations of hyperparameter sequences. If training on each set takes 6 minutes, exhaustive training on all potential combinations of the hyperparameters to find the optimal values will take almost 3000 years. Expert knowledge or random selection are some alternate options, but they are not scalable or consistently reliable. Metaheuristics such as evolutionary algorithms are a great choice for solving combinatorial optimization problems like hyperparameter optimization. While other researchers have used evolutionary algorithms such as standard implementation of genetic algorithm (GA), we introduce additional nature-inspired enhancements to GA for better exploration of the hyperparameter solution space to optimize the DNN architecture. The training is complemented with Monte Carlo based variance reduction method called importance sampling. We demonstrate that these fine-tunings result in improvements in the network accuracy on MNIST and CIFAR-10 datasets that outperforms standard use of genetic algorithm.