Sparse tensors appear frequently in distributed deep learning, either as a direct artifact of the deep neural network's gradients, or as a result of an explicit sparsification process. Existing communication primitives are agnostic to the peculiarities of deep learning; consequently, they impose unnecessary communication overhead. This paper introduces DeepReduce, a versatile framework for the compressed communication of sparse tensors, tailored for distributed deep learning. DeepReduce decomposes sparse tensors in two sets, values and indices, and allows both independent and combined compression of these sets. We support a variety of common compressors, such as Deflate for values, or run-length encoding for indices. We also propose two novel compression schemes that achieve superior results: curve fitting-based for values and bloom filter-based for indices. DeepReduce is orthogonal to existing gradient sparsifiers and can be applied in conjunction with them, transparently to the end-user, to significantly lower the communication overhead. As proof of concept, we implement our approach on Tensorflow and PyTorch. Our experiments with large real models demonstrate that DeepReduce transmits fewer data and imposes lower computational overhead than existing methods, without affecting the training accuracy.
Increasing the batch size of a deep learning model is a challenging task. Although it might help in utilizing full available system memory during training phase of a model, it results in significant loss of test accuracy most often. LARS solved this issue by introducing an adaptive learning rate for each layer of a deep learning model. However, there are doubts on how popular distributed machine learning systems such as SystemML or MLlib will perform with this optimizer. In this work, we apply LARS optimizer to a deep learning model implemented using SystemML.We perform experiments with various batch sizes and compare the performance of LARS optimizer with \textit{Stochastic Gradient Descent}. Our experimental results show that LARS optimizer performs significantly better than Stochastic Gradient Descent for large batch sizes even with the distributed machine learning framework, SystemML.
There has been an increasing interest in harnessing deep learning to tackle combinatorial optimization (CO) problems in recent years. Typical CO deep learning approaches leverage the problem structure in the model architecture. Nevertheless, the model selection is still mainly based on the conventional machine learning setting. Due to the discrete nature of CO problems, a single model is unlikely to learn the problem entirely. We introduce ZTop, which stands for Zero Training Overhead Portfolio, a simple yet effective model selection and ensemble mechanism for learning to solve combinatorial problems. ZTop is inspired by algorithm portfolios, a popular CO ensembling strategy, particularly restart portfolios, which periodically restart a randomized CO algorithm, de facto exploring the search space with different heuristics. We have observed that well-trained models acquired in the same training trajectory, with similar top validation performance, perform well on very different validation instances. Following this observation, ZTop ensembles a set of well-trained models, each providing a unique heuristic with zero training overhead, and applies them, sequentially or in parallel, to solve the test instances. We show how ZTopping, i.e., using a ZTop ensemble strategy with a given deep learning approach, can significantly improve the performance of the current state-of-the-art deep learning approaches on three prototypical CO domains, the hardest unique-solution Sudoku instances, challenging routing problems, and the graph maximum cut problem, as well as on multi-label classification, a machine learning task with a large combinatorial label space.
Many minimally invasive interventional procedures still rely on 2D fluoroscopic imaging. Generating a patient-specific 3D model from these X-ray projection data would allow to improve the procedural workflow, e.g. by providing assistance functions such as automatic positioning. To accomplish this, two things are required. First, a statistical human shape model of the human anatomy and second, a differentiable X-ray renderer. In this work, we propose a differentiable renderer by deriving the distance travelled by a ray inside mesh structures to generate a distance map. To demonstrate its functioning, we use it for simulating X-ray images from human shape models. Then we show its application by solving the inverse problem, namely reconstructing 3D models from real 2D fluoroscopy images of the pelvis, which is an ideal anatomical structure for patient registration. This is accomplished by an iterative optimization strategy using gradient descent. With the majority of the pelvis being in the fluoroscopic field of view, we achieve a mean Hausdorff distance of 30 mm between the reconstructed model and the ground truth segmentation.
We introduce Ivy, a templated Deep Learning (DL) framework which abstracts existing DL frameworks. Ivy unifies the core functions of these frameworks to exhibit consistent call signatures, syntax and input-output behaviour. New high-level framework-agnostic functions and classes, which are usable alongside framework-specific code, can then be implemented as compositions of the unified low-level Ivy functions. Ivy currently supports TensorFlow, PyTorch, MXNet, Jax and NumPy. We also release four pure-Ivy libraries for mechanics, 3D vision, robotics, and differentiable environments. Through our evaluations, we show that Ivy can significantly reduce lines of code with a runtime overhead of less than 1% in most cases. We welcome developers to join the Ivy community by writing their own functions, layers and libraries in Ivy, maximizing their audience and helping to accelerate DL research through inter-framework codebases. More information can be found at https://ivy-dl.org.
Object detection in natural scenes can be a challenging task. In many real-life situations, the visible spectrum is not suitable for traditional computer vision tasks. Moving outside the visible spectrum range, such as the thermal spectrum or the near-infrared (NIR) images, is much more beneficial in low visibility conditions, NIR images are very helpful for understanding the object's material quality. In this work, we have taken images with both the Thermal and NIR spectrum for the object detection task. As multi-spectral data with both Thermal and NIR is not available for the detection task, we needed to collect data ourselves. Data collection is a time-consuming process, and we faced many obstacles that we had to overcome. We train the YOLO v3 network from scratch to detect an object from multi-spectral images. Also, to avoid overfitting, we have done data augmentation and tune hyperparameters.
Data is evolving with the rapid progress of population and communication for various types of devices such as networks, cloud computing, Internet of Things (IoT), actuators, and sensors. The increment of data and communication content goes with the equivalence of velocity, speed, size, and value to provide the useful and meaningful knowledge that helps to solve the future challenging tasks and latest issues. Besides, multicriteria based decision making is one of the key issues to solve for various issues related to the alternative effects in big data analysis. It tends to find a solution based on the latest machine learning techniques that include algorithms like decision making and deep learning mechanism based on multicriteria in providing insights to big data. On the other hand, the derivations are made for it to go with the approximations to increase the duality of runtime and improve the entire system's potentiality and efficacy. In essence, several fields, including business, agriculture, information technology, and computer science, use deep learning and multicriteria-based decision-making problems. This paper aims to provide various applications that involve the concepts of deep learning techniques and exploiting the multicriteria approaches for issues that are facing in big data analytics by proposing new studies with the fusion approaches of data-driven techniques.
From ancient to modern times, acoustic structures have been used to control the propagation of acoustic waves. However, the design of the acoustic structures has remained widely a time-consuming and computational resource-consuming iterative process. In recent years, Deep Learning has attracted unprecedented attention for its ability to tackle hard problems with huge datasets, which has achieved state-of-the-art results in various tasks. In this work, an acoustic structure design method is proposed based on deep learning. Taking the design of multi-order Helmholtz resonator for instance, we experimentally demonstrate the effectiveness of the proposed method. Our method is not only able to give a very accurate prediction of the geometry of the acoustic structures with multiple strong-coupling parameters, but also capable of improving the performance of evolutionary approaches in optimization for a desired property. Compared with the conventional numerical methods, our method is more efficient, universal and automatic, which has a wide range of potential applications, such as speech enhancement, sound absorption and insulation.
A decision support system relies on frequent re-solving of similar problem instances. While the general structure remains the same in corresponding applications, the input parameters are updated on a regular basis. We propose a generative neural network design for learning integer decision variables of mixed-integer linear programming (MILP) formulations of these problems. We utilise a deep neural network discriminator and a MILP solver as our oracle to train our generative neural network. In this article, we present the results of our design applied to the transient gas optimisation problem. With the trained network we produce a feasible solution in 2.5s, use it as a warm-start solution, and thereby decrease global optimal solution solve time by 60.5%.
One of the major limitations of deep learning models is that they face catastrophic forgetting in an incremental learning scenario. There have been several approaches proposed to tackle the problem of incremental learning. Most of these methods are based on knowledge distillation and do not adequately utilize the information provided by older task models, such as uncertainty estimation in predictions. The predictive uncertainty provides the distributional information can be applied to mitigate catastrophic forgetting in a deep learning framework. In the proposed work, we consider a Bayesian formulation to obtain the data and model uncertainties. We also incorporate self-attention framework to address the incremental learning problem. We define distillation losses in terms of aleatoric uncertainty and self-attention. In the proposed work, we investigate different ablation analyses on these losses. Furthermore, we are able to obtain better results in terms of accuracy on standard benchmarks.
Multi-scale biomedical knowledge networks are expanding with emerging experimental technologies that generates multi-scale biomedical big data. Link prediction is increasingly used especially in bipartite biomedical networks to identify hidden biological interactions and relationshipts between key entities such as compounds, targets, gene and diseases. We propose a Graph Neural Networks (GNN) method, namely Graph Pair based Link Prediction model (GPLP), for predicting biomedical network links simply based on their topological interaction information. In GPLP, 1-hop subgraphs extracted from known network interaction matrix is learnt to predict missing links. To evaluate our method, three heterogeneous biomedical networks were used, i.e. Drug-Target Interaction network (DTI), Compound-Protein Interaction network (CPI) from NIH Tox21, and Compound-Virus Inhibition network (CVI). Our proposed GPLP method significantly outperforms over the state-of-the-art baselines. In addition, different network incompleteness is analysed with our devised protocol, and we also design an effective approach to improve the model robustness towards incomplete networks. Our method demonstrates the potential applications in other biomedical networks.
B-mode ultrasound imaging is a popular medical imaging technique. Like other image processing tasks, deep learning has been used for analysis of B-mode ultrasound images in the last few years. However, training deep learning models requires large labeled datasets, which is often unavailable for ultrasound images. The lack of large labeled data is a bottleneck for the use of deep learning in ultrasound image analysis. To overcome this challenge, in this work we exploit Auxiliary Classifier Generative Adversarial Network (ACGAN) that combines the benefits of data augmentation and transfer learning in the same framework. We conduct experiment on a dataset of breast ultrasound images that shows the effectiveness of the proposed approach.
Federated Learning marks a turning point in the implementation of decentralized machine learning (especially deep learning) for wireless devices by protecting users' privacy and safeguarding raw data from third-party access. It assigns the learning process independently to each client. First, clients locally train a machine learning model based on local data. Next, clients transfer local updates of model weights and biases (training data) to a server. Then, the server aggregates updates (received from clients) to create a global learning model. However, the continuous transfer between clients and the server increases communication costs and is inefficient from a resource utilization perspective due to the large number of parameters (weights and biases) used by deep learning models. The cost of communication becomes a greater concern when the number of contributing clients and communication rounds increases. In this work, we propose a novel framework, FedZip, that significantly decreases the size of updates while transferring weights from the deep learning model between clients and their servers. FedZip implements Top-z sparsification, uses quantization with clustering, and implements compression with three different encoding methods. FedZip outperforms state-of-the-art compression frameworks and reaches compression rates up to 1085x, and preserves up to 99% of bandwidth and 99% of energy for clients during communication.
Deep learning (DL) has gained much attention and become increasingly popular in modern data science. Computer scientists led the way in developing deep learning techniques, so the ideas and perspectives can seem alien to statisticians. Nonetheless, it is important that statisticians become involved -- many of our students need this expertise for their careers. In this paper, developed as part of a program on DL held at the Statistical and Applied Mathematical Sciences Institute, we address this culture gap and provide tips on how to teach deep learning to statistics graduate students. After some background, we list ways in which DL and statistical perspectives differ, provide a recommended syllabus that evolved from teaching two iterations of a DL graduate course, offer examples of suggested homework assignments, give an annotated list of teaching resources, and discuss DL in the context of two research areas.
This volume contains the papers accepted at the first DATE Friday Workshop on System-level Design Methods for Deep Learning on Heterogeneous Architectures (SLOHA 2021), held virtually on February 5, 2021. SLOHA 2021 was co-located with the Conference on Design, Automation and Test in Europe (DATE).
For the sake of recognizing and classifying textile defects, deep learning-based methods have been proposed and achieved remarkable success in single-label textile images. However, detecting multi-label defects in a textile image remains challenging due to the coexistence of multiple defects and small-size defects. To address these challenges, a multi-level, multi-attentional deep learning network (MLMA-Net) is proposed and built to 1) increase the feature representation ability to detect small-size defects; 2) generate a discriminative representation that maximizes the capability of attending the defect status, which leverages higher-resolution feature maps for multiple defects. Moreover, a multi-label object detection dataset (DHU-ML1000) in textile defect images is built to verify the performance of the proposed model. The results demonstrate that the network extracts more distinctive features and has better performance than the state-of-the-art approaches on the real-world industrial dataset.
Retinal degenerative diseases cause profound visual impairment in more than 10 million people worldwide, and retinal prostheses are being developed to restore vision to these individuals. Analogous to cochlear implants, these devices electrically stimulate surviving retinal cells to evoke visual percepts (phosphenes). However, the quality of current prosthetic vision is still rudimentary. Rather than aiming to restore "natural" vision, there is potential merit in borrowing state-of-the-art computer vision algorithms as image processing techniques to maximize the usefulness of prosthetic vision. Here we combine deep learning--based scene simplification strategies with a psychophysically validated computational model of the retina to generate realistic predictions of simulated prosthetic vision, and measure their ability to support scene understanding of sighted subjects (virtual patients) in a variety of outdoor scenarios. We show that object segmentation may better support scene understanding than models based on visual saliency and monocular depth estimation. In addition, we highlight the importance of basing theoretical predictions on biologically realistic models of phosphene shape. Overall, this work has the potential to drastically improve the utility of prosthetic vision for people blinded from retinal degenerative diseases.
In industry NLP application, our manually labeled data has a certain number of noisy data. We present a simple method to find the noisy data and re-label them manually, meanwhile we collect the correction information. Then we present novel method to incorporate the human correction information into deep learning model. Human know how to correct noisy data. So the correction information can be inject into deep learning model. We do the experiment on our own text classification dataset, which is manually labeled, because we re-label the noisy data in our dataset for our industry application. The experiment result shows that our method improve the classification accuracy from 91.7% to 92.5%. The 91.7% accuracy is trained on the corrected dataset, which improve the baseline from 83.3% to 91.7%.
Numerical weather forecasting using high-resolution physical models often requires extensive computational resources on supercomputers, which diminishes their wide usage in most real-life applications. As a remedy, applying deep learning methods has revealed innovative solutions within this field. To this end, we introduce a novel deep learning architecture for forecasting high-resolution spatio-temporal weather data. Our approach extends the conventional encoder-decoder structure by integrating Convolutional Long-short Term Memory and Convolutional Neural Networks. In addition, we incorporate attention and context matcher mechanisms into the model architecture. Our Weather Model achieves significant performance improvements compared to baseline deep learning models, including ConvLSTM, TrajGRU, and U-Net. Our experimental evaluation involves high-scale, real-world benchmark numerical weather datasets, namely the ERA5 hourly dataset on pressure levels and WeatherBench. Our results demonstrate substantial improvements in identifying spatial and temporal correlations with attention matrices focusing on distinct parts of the input series to model atmospheric circulations. We also compare our model with high-resolution physical models using the benchmark metrics and show that our Weather Model is accurate and easy to interpret.
Fractures occur in the shoulder area, which has a wider range of motion than other joints in the body, for various reasons. To diagnose these fractures, data gathered from Xradiation (X-ray), magnetic resonance imaging (MRI), or computed tomography (CT) are used. This study aims to help physicians by classifying shoulder images taken from X-ray devices as fracture / non-fracture with artificial intelligence. For this purpose, the performances of 26 deep learning-based pretrained models in the detection of shoulder fractures were evaluated on the musculoskeletal radiographs (MURA) dataset, and two ensemble learning models (EL1 and EL2) were developed. The pretrained models used are ResNet, ResNeXt, DenseNet, VGG, Inception, MobileNet, and their spinal fully connected (Spinal FC) versions. In the EL1 and EL2 models developed using pretrained models with the best performance, test accuracy was 0.8455,0.8472, Cohens kappa was 0.6907, 0.6942 and the area that was related with fracture class under the receiver operating characteristic (ROC) curve (AUC) was 0.8862,0.8695. As a result of 28 different classifications in total, the highest test accuracy and Cohens kappa values were obtained in the EL2 model, and the highest AUC value was obtained in the EL1 model.