Learning by Association — A Versatile Semi-Supervised Training Method for Neural Networks

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.

Recently, how to exploit unlabeled data for training face recognition models has been attracting increasing attention. However, few works consider the unlabeled shallow data <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> in real-world scenarios. The existing semi-supervised face recognition methods that focus on generating pseudo labels or minimizing softmax classification probabilities of the unlabeled data do not work very well on the unlabeled shallow data. It is still a challenge on how to effectively utilize the unlabeled shallow face data to improve the performance of face recognition. In this paper, we propose a novel face recognition method, named VirFace, to effectively exploit the unlabeled shallow data for face recognition. VirFace consists of VirClass and VirInstance. Specifically, VirClass enlarges the inter-class distance by injecting the unlabeled data as new identities, while VirInstance produces virtual instances sampled from the learned distribution of each identity to further enlarge the inter-class distance. To the best of our knowledge, we are the first to tackle the problem of unlabeled shallow face data. Extensive experiments have been conducted on both the small- and large-scale datasets, e.g. LFW and IJB-C, etc, demonstrating the superiority of the proposed method.

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.

Error Correcting Output Codes, ECOC, is an output representation method capable of discovering some of the errors produced in classification tasks. This paper describes the application of ECOC to the training of feed forward neural networks, FFNN, for improving the overall accuracy of classification systems. Indeed, to improve the generalization of FFNN classifiers, this paper proposes an ECOC-Based training method for Neural Networks that use ECOC as the output representation, and adopts the traditional Back-Propagation algorithm, BP, to adjust weights of the network. Experimental results for face recognition problem on Yale database demonstrate the effectiveness of our method. With a rejection scheme defined by a simple robustness rate, high reliability is achieved in this application.

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.

In this paper we search for an optimal strategy for semi-supervised Deep Neural Network (DNN) training. We assume that a small part of the data is transcribed, while the majority of the data is untranscribed. We explore self-training strategies with data selection based on both the utterance-level and frame-level confidences. Further on, we study the interactions between semi-supervised frame-discriminative training and sequence-discriminative sMBR training. We found it beneficial to reduce the disproportion in amounts of transcribed and untranscribed data by including the transcribed data several times, as well as to do a frame-selection based on per-frame confidences derived from confusion in a lattice. For the experiments, we used the Limited language pack condition for the Surprise language task (Vietnamese) from the IARPA Babel program. The absolute Word Error Rate (WER) improvement for frame cross-entropy training is 2.2%, this corresponds to WER recovery of 36% when compared to the identical system, where the DNN is built on the fully transcribed data.

ADANOISE: Training neural networks with adaptive noise for imbalanced data classification

We describe IBM's roadmap for Neuromorphic Technologies to drive next-generation cognitive computing, ranging from nanodevice-based hardware for accelerating well-known supervised-learning algorithms (which happen to rely on static, labeled data), to emerging, biologically-inspired algorithms capable of learning from temporal, unlabeled data. The various hardware-centric neuromorphic projects currently underway at IBM Research will be surveyed, with a focus on the use of Non-Volatile Memory (NVM) for on-chip acceleration of the training of Deep Neural Networks (DNNs).

The recent rapid success of deep convolutional neural networks (CNN) on many computer vision tasks largely benefits from the well-annotated Pascal VOC, ImageNet, and MS COCO datasets. However, it is challenging to get ImageNet-like annotations (1000 classes) in the medical imaging domain due to the lack of clinical training in the lay crowdsourcing community. We address this problem by presenting a semi-supervised training method for neural networks with true-class and pseudo-class (un-annotated class) labels on partially annotated training data. The true-class labels are supervised annotations from clinical professionals. The pseudo-class labels are unsupervised clustering of un-annotated data. Our method rests upon the hypothesis of better coherent annotations with discriminative classes leading to better trained CNN models. We validated our method on extra-coronary calcification detection in low dose CT scans. The CNN trained with true-class and 10 pseudo-classes achieved an 85.0% sensitivity at 10 false positives per patient (0.3 false positive per slice), which significantly outperformed the CNN trained with true-class only (sensitivity =56.0% at 10 false positives per patient).

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.

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

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.

We characterize and remedy a failure mode that may arise from multi-scale dynamics with scale imbalances during training of deep neural networks, such as physics informed neural networks (PINNs). PINNs are popular machine-learning templates that allow for seamless integration of physical equation models with data. Their training amounts to solving an optimization problem over a weighted sum of data-fidelity and equation-fidelity objectives. Conflicts between objectives can arise from scale imbalances, heteroscedasticity in the data, stiffness of the physical equation, or from catastrophic interference during sequential training. We explain the training pathology arising from this and propose a simple yet effective inverse Dirichlet weighting strategy to alleviate the issue. We compare with Sobolev training of neural networks, providing the baseline of analytically ε-optimal training. We demonstrate the effectiveness of inverse Dirichlet weighting in various applications, including a multi-scale model of active turbulence, where we show orders of magnitude improvement in accuracy and convergence over conventional PINN training. For inverse modeling using sequential training, we find that inverse Dirichlet weighting protects a PINN against catastrophic forgetting.