OVID-19 is a world-wide disease that has been declared as a pandemic by the World Health Organization. Computer Tomography (CT) imaging of the chest seems to be a valid diagnosis tool to detect COVID-19 promptly and to control the spread of the disease. Deep Learning has been extensively used in medical imaging and convolutional neural networks (CNNs) have been also used for classification of CT images. We propose a light CNN design based on the model of the SqueezeNet, for the efficient discrimination of COVID-19 CT images with other CT images (community-acquired pneumonia and/or healthy images). On the tested datasets, the proposed modified SqueezeNet CNN achieved 83.00\% of accuracy, 85.00\% of sensitivity, 81.00\% of specificity, 81.73\% of precision and 0.8333 of F1Score in a very efficient way (7.81 seconds medium-end laptot without GPU acceleration). Besides performance, the average classification time is very competitive with respect to more complex CNN designs, thus allowing its usability also on medium power computers. In the next future we aim at improving the performances of the method along two directions: 1) by increasing the training dataset (as soon as other CT images will be available); 2) by introducing an efficient pre-processing strategy.
In this paper, we introduce a novel single shot approach for 6D object pose estimation of rigid objects based on depth images. For this purpose, a fully convolutional neural network is employed, where the 3D input data is spatially discretized and pose estimation is considered as a regression task that is solved locally on the resulting volume elements. With 65 fps on a GPU, our Object Pose Network (OP-Net) is extremely fast, is optimized end-to-end, and estimates the 6D pose of multiple objects in the image simultaneously. Our approach does not require manually 6D pose-annotated real-world datasets and transfers to the real world, although being entirely trained on synthetic data. The proposed method is evaluated on public benchmark datasets, where we can demonstrate that state-of-the-art methods are significantly outperformed.
Lidar based 3D object detection and classification tasks are essential for automated driving(AD). A Lidar sensor can provide the 3D point coud data reconstruction of the surrounding environment. But the detection in 3D point cloud still needs a strong algorithmic challenge. This paper consists of three parts.(1)Lidar-camera calib. (2)YOLO, based detection and PointCloud extraction, (3) k-means based point cloud segmentation. In our research, Camera can capture the image to make the Real-time 2D Object Detection by using YOLO, I transfer the bounding box to node whose function is making 3d object detection on point cloud data from Lidar. By comparing whether 2D coordinate transferred from the 3D point is in the object bounding box or not, and doing a k-means clustering can achieve High-speed 3D object recognition function in GPU.
Transformer has become ubiquitous in natural language processing (e.g., machine translation, question answering); however, it requires enormous amount of computations to achieve high performance, which makes it not suitable for mobile applications that are tightly constrained by the hardware resources and battery. In this paper, we present an efficient mobile NLP architecture, Lite Transformer to facilitate deploying mobile NLP applications on edge devices. The key primitive is the Long-Short Range Attention (LSRA), where one group of heads specializes in the local context modeling (by convolution) while another group specializes in the long-distance relationship modeling (by attention). Such specialization brings consistent improvement over the vanilla transformer on three well-established language tasks: machine translation, abstractive summarization, and language modeling. Under constrained resources (500M/100M MACs), Lite Transformer outperforms transformer on WMT'14 English-French by 1.2/1.7 BLEU, respectively. Lite Transformer reduces the computation of transformer base model by 2.5x with 0.3 BLEU score degradation. Combining with pruning and quantization, we further compressed the model size of Lite Transformer by 18.2x. For language modeling, Lite Transformer achieves 1.8 lower perplexity than the transformer at around 500M MACs. Notably, Lite Transformer outperforms the AutoML-based Evolved Transformer by 0.5 higher BLEU for the mobile NLP setting without the costly architecture search that requires more than 250 GPU years. Code has been made available at https://github.com/mit-han-lab/lite-transformer.
We leverage deep learning techniques to jointly denoise and super-resolve biomedical images acquired with fluorescence microscopy. We develop a deep learning algorithm based on the networks and method described in the recent W2S paper to solve a joint denoising and super-resolution problem. Specifically, we address the restoration of SIM images from widefield images. Our TensorFlow model is trained on the W2S dataset of cell images and is made accessible online in this repository: https://github.com/mchatton/w2s-tensorflow. On test images, the model shows a visually-convincing denoising and increases the resolution by a factor of two compared to the input image. For a 512 $\times$ 512 image, the inference takes less than 1 second on a Titan X GPU and about 15 seconds on a common CPU. We further present the results of different variations of losses used in training.
The KeOps library provides a fast and memory-efficient GPU support for tensors whose entries are given by a mathematical formula, such as kernel and distance matrices. KeOps alleviates the major bottleneck of tensor-centric libraries for kernel and geometric applications: memory consumption. It also supports automatic differentiation and outperforms standard GPU baselines, including PyTorch CUDA tensors or the Halide and TVM libraries. KeOps combines optimized C++/CUDA schemes with binders for high-level languages: Python (Numpy and PyTorch), Matlab and GNU R. As a result, high-level "quadratic" codes can now scale up to large data sets with millions of samples processed in seconds. KeOps brings graphics-like performances for kernel methods and is freely available on standard repositories (PyPi, CRAN). To showcase its versatility, we provide tutorials in a wide range of settings online at \url{www.kernel-operations.io}.
We present a neural network TTNet aimed at real-time processing of high-resolution table tennis videos, providing both temporal (events spotting) and spatial (ball detection and semantic segmentation) data. This approach gives core information for reasoning score updates by an auto-referee system. We also publish a multi-task dataset OpenTTGames with videos of table tennis games in 120 fps labeled with events, semantic segmentation masks, and ball coordinates for evaluation of multi-task approaches, primarily oriented on spotting of quick events and small objects tracking. TTNet demonstrated 97.0% accuracy in game events spotting along with 2 pixels RMSE in ball detection with 97.5% accuracy on the test part of the presented dataset. The proposed network allows the processing of downscaled full HD videos with inference time below 6 ms per input tensor on a machine with a single consumer-grade GPU. Thus, we are contributing to the development of real-time multi-task deep learning applications and presenting approach, which is potentially capable of substituting manual data collection by sports scouts, providing support for referees' decision-making, and gathering extra information about the game process.
Recent advancements in Neural Architecture Search(NAS) resulted in finding new state-of-the-art Artificial Neural Network (ANN) solutions for tasks like image classification, object detection, or semantic segmentation without substantial human supervision. In this paper, we focus on exploring NAS for a dense prediction task that is image denoising. Due to a costly training procedure, most NAS solutions for image enhancement rely on reinforcement learning or evolutionary algorithm exploration, which usually take weeks (or even months) to train. Therefore, we introduce a new efficient implementation of various superkernel techniques that enable fast (6-8 RTX2080 GPU hours) single-shot training of models for dense predictions. We demonstrate the effectiveness of our method on the SIDD+ benchmark for image denoising.
The widely-adopted practice is to train deep learning models with specialized hardware accelerators, e.g., GPUs or TPUs, due to their superior performance on linear algebra operations. However, this strategy does not employ effectively the extensive CPU and memory resources -- which are used only for preprocessing, data transfer, and scheduling -- available by default on the accelerated servers. In this paper, we study training algorithms for deep learning on heterogeneous CPU+GPU architectures. Our two-fold objective -- maximize convergence rate and resource utilization simultaneously -- makes the problem challenging. In order to allow for a principled exploration of the design space, we first introduce a generic deep learning framework that exploits the difference in computational power and memory hierarchy between CPU and GPU through asynchronous message passing. Based on insights gained through experimentation with the framework, we design two heterogeneous asynchronous stochastic gradient descent (SGD) algorithms. The first algorithm -- CPU+GPU Hogbatch -- combines small batches on CPU with large batches on GPU in order to maximize the utilization of both resources. However, this generates an unbalanced model update distribution which hinders the statistical convergence. The second algorithm -- Adaptive Hogbatch -- assigns batches with continuously evolving size based on the relative speed of CPU and GPU. This balances the model updates ratio at the expense of a customizable decrease in utilization. We show that the implementation of these algorithms in the proposed CPU+GPU framework achieves both faster convergence and higher resource utilization than TensorFlow on several real datasets and on two computing architectures -- an on-premises server and a cloud instance.
Dense, discrete Graphical Models with pairwise potentials are a powerful class of models which are employed in state-of-the-art computer vision and bio-imaging applications. This work introduces a new MAP-solver, based on the popular Dual Block-Coordinate Ascent principle. Surprisingly, by making a small change to the low-performing solver, the Max Product Linear Programming (MPLP) algorithm, we derive the new solver MPLP++ that significantly outperforms all existing solvers by a large margin, including the state-of-the-art solver Tree-Reweighted Sequential (TRWS) message-passing algorithm. Additionally, our solver is highly parallel, in contrast to TRWS, which gives a further boost in performance with the proposed GPU and multi-thread CPU implementations. We verify the superiority of our algorithm on dense problems from publicly available benchmarks, as well, as a new benchmark for 6D Object Pose estimation. We also provide an ablation study with respect to graph density.
Spectral methods have been the mainstay in several domains such as machine learning and scientific computing. They involve finding a certain kind of spectral decomposition to obtain basis functions that can capture important structures for the problem at hand. The most common spectral method is the principal component analysis (PCA). It utilizes the top eigenvectors of the data covariance matrix, e.g. to carry out dimensionality reduction. This data pre-processing step is often effective in separating signal from noise. PCA and other spectral techniques applied to matrices have several limitations. By limiting to only pairwise moments, they are effectively making a Gaussian approximation on the underlying data and fail on data with hidden variables which lead to non-Gaussianity. However, in most data sets, there are latent effects that cannot be directly observed, e.g., topics in a document corpus, or underlying causes of a disease. By extending the spectral decomposition methods to higher order moments, we demonstrate the ability to learn a wide range of latent variable models efficiently. Higher-order moments can be represented by tensors, and intuitively, they can encode more information than just pairwise moment matrices. More crucially, tensor decomposition can pick up latent effects that are missed by matrix methods, e.g. uniquely identify non-orthogonal components. Exploiting these aspects turns out to be fruitful for provable unsupervised learning of a wide range of latent variable models. We also outline the computational techniques to design efficient tensor decomposition methods. We introduce Tensorly, which has a simple python interface for expressing tensor operations. It has a flexible back-end system supporting NumPy, PyTorch, TensorFlow and MXNet amongst others, allowing multi-GPU and CPU operations and seamless integration with deep-learning functionalities.
GPUs are a key enabler of the revolution in machine learning and high performance computing, functioning as de facto co-processors to accelerate large-scale computation. As the programming stack and tool support have matured, GPUs have also become accessible to programmers, who may lack detailed knowledge of the underlying architecture and fail to fully leverage the GPU's computation power. GEVO (Gpu optimization using EVOlutionary computation) is a tool for automatically discovering optimization opportunities and tuning the performance of GPU kernels in the LLVM representation. GEVO uses population-based search to find edits to GPU code compiled to LLVM-IR and improves performance on desired criteria while retaining required functionality. We demonstrate that GEVO improves the execution time of the GPU programs in the Rodinia benchmark suite and the machine learning models, SVM and ResNet18, on NVIDIA Tesla P100. For the Rodinia benchmarks, GEVO improves GPU kernel runtime performance by an average of 49.48% and by as much as 412% over the fully compiler-optimized baseline. If kernel output accuracy is relaxed to tolerate up to 1% error, GEVO can find kernel variants that outperform the baseline version by an average of 51.08%. For the machine learning workloads, GEVO achieves kernel performance improvement for SVM on the MNIST handwriting recognition (3.24X) and the a9a income prediction (2.93X) datasets with no loss of model accuracy. GEVO achieves 1.79X kernel performance improvement on image classification using ResNet18/CIFAR-10, with less than 1% model accuracy reduction.
1. Joint Species Distribution models (JSDMs) explain spatial variation in community composition by contributions of the environment, biotic associations, and possibly spatially structured residual covariance. They show great promise as a general analytical framework for community ecology and macroecology, but current JSDMs, even when approximated by latent variables, scale poorly on large datasets, limiting their usefulness for currently emerging big (e.g., metabarcoding and metagenomics) community datasets. 2. Here, we present a novel, more scalable JSDM (sjSDM) that circumvents the need to use latent variables by using a Monte-Carlo integration of the joint JSDM likelihood and allows flexible elastic net regularization on all model components. We implemented sjSDM in PyTorch, a modern machine learning framework that can make use of CPU and GPU calculations. Using simulated communities with known species-species associations and different number of species and sites, we compare sjSDM with state-of-the-art JSDM implementations to determine computational runtimes and accuracy of the inferred species-species and species-environmental associations. 3. We find that sjSDM is orders of magnitude faster than existing JSDM algorithms (even when run on the CPU) and can be scaled to very large datasets. Despite the dramatically improved speed, sjSDM produces more accurate estimates of species association structures than alternative JSDM implementations. We demonstrate the applicability of sjSDM to big community data using eDNA case study with thousands of fungi operational taxonomic units (OTU). 4. Our sjSDM approach makes the analysis of JSDMs to large community datasets with hundreds or thousands of species possible, substantially extending the applicability of JSDMs in ecology. We provide our method in an R package to facilitate its applicability for practical data analysis.
Conventional neural accelerators rely on isolated self-sufficient functional units that perform an atomic operation while communicating the results through an operand delivery-aggregation logic. Each single unit processes all the bits of their operands atomically and produce all the bits of the results in isolation. This paper explores a different design style, where each unit is only responsible for a slice of the bit-level operations to interleave and combine the benefits of bit-level parallelism with the abundant data-level parallelism in deep neural networks. A dynamic collection of these units cooperate at runtime to generate bits of the results, collectively. Such cooperation requires extracting new grouping between the bits, which is only possible if the operands and operations are vectorizable. The abundance of Data Level Parallelism and mostly repeated execution patterns, provides a unique opportunity to define and leverage this new dimension of Bit-Parallel Vector Composability. This design intersperses bit parallelism within data-level parallelism and dynamically interweaves the two together. As such, the building block of our neural accelerator is a Composable Vector Unit that is a collection of Narrower-Bitwidth Vector Engines, which are dynamically composed or decomposed at the bit granularity. Using six diverse CNN and LSTM deep networks, we evaluate this design style across four design points: with and without algorithmic bitwidth heterogeneity and with and without availability of a high-bandwidth off-chip memory. Across these four design points, Bit-Parallel Vector Composability brings (1.4x to 3.5x) speedup and (1.1x to 2.7x) energy reduction. We also comprehensively compare our design style to the Nvidia RTX 2080 TI GPU, which also supports INT-4 execution. The benefits range between 28.0x and 33.7x improvement in Performance-per-Watt.
Deep learning networks are being developed in every stage of the MRI workflow and have provided state-of-the-art results. However, this has come at the cost of increased computation requirement and storage. Hence, replacing the networks with compact models at various stages in the MRI workflow can significantly reduce the required storage space and provide considerable speedup. In computer vision, knowledge distillation is a commonly used method for model compression. In our work, we propose a knowledge distillation (KD) framework for the image to image problems in the MRI workflow in order to develop compact, low-parameter models without a significant drop in performance. We propose a combination of the attention-based feature distillation method and imitation loss and demonstrate its effectiveness on the popular MRI reconstruction architecture, DC-CNN. We conduct extensive experiments using Cardiac, Brain, and Knee MRI datasets for 4x, 5x and 8x accelerations. We observed that the student network trained with the assistance of the teacher using our proposed KD framework provided significant improvement over the student network trained without assistance across all the datasets and acceleration factors. Specifically, for the Knee dataset, the student network achieves $65\%$ parameter reduction, 2x faster CPU running time, and 1.5x faster GPU running time compared to the teacher. Furthermore, we compare our attention-based feature distillation method with other feature distillation methods. We also conduct an ablative study to understand the significance of attention-based distillation and imitation loss. We also extend our KD framework for MRI super-resolution and show encouraging results.
Variational inference methods have been shown to lead to significant improvements in the computational efficiency of approximate Bayesian inference in mixed multinomial logit models when compared to standard Markov-chain Monte Carlo (MCMC) methods without compromising accuracy. However, despite their demonstrated efficiency gains, existing methods still suffer from important limitations that prevent them to scale to very large datasets, while providing the flexibility to allow for rich prior distributions and to capture complex posterior distributions. In this paper, we propose an Amortized Variational Inference approach that leverages stochastic backpropagation, automatic differentiation and GPU-accelerated computation, for effectively scaling Bayesian inference in Mixed Multinomial Logit models to very large datasets. Moreover, we show how normalizing flows can be used to increase the flexibility of the variational posterior approximations. Through an extensive simulation study, we empirically show that the proposed approach is able to achieve computational speedups of multiple orders of magnitude over traditional MSLE and MCMC approaches for large datasets without compromising estimation accuracy.
Pairwise comparison data arise in many domains with subjective assessment experiments, for example in image and video quality assessment. In these experiments observers are asked to express a preference between two conditions. However, many pairwise comparison protocols require a large number of comparisons to infer accurate scores, which may be unfeasible when each comparison is time-consuming (e.g. videos) or expensive (e.g. medical imaging). This motivates the use of an active sampling algorithm that chooses only the most informative pairs for comparison. In this paper we propose ASAP, an active sampling algorithm based on approximate message passing and expected information gain maximization. Unlike most existing methods, which rely on partial updates of the posterior distribution, we are able to perform full updates and therefore much improve the accuracy of the inferred scores. The algorithm relies on three techniques for reducing computational cost: inference based on approximate message passing, selective evaluations of the information gain, and selecting pairs in a batch that forms a minimum spanning tree of the inverse of information gain. We demonstrate, with real and synthetic data, that ASAP offers the highest accuracy of inferred scores compared to the existing methods. We also provide an open-source GPU implementation of ASAP for large-scale experiments.
LIDAR (light detection and ranging) is an optical remote-sensing technique that measures the distance between sensor and object, and the reflected energy from the object. Over the years, LIDAR data has been used as the primary source of Digital Elevation Models (DEMs). DEMs have been used in a variety of applications like road extraction, hydrological modeling, flood mapping, and surface analysis. A number of studies in flooding suggest the usage of high-resolution DEMs as inputs in the applications improve the overall reliability and accuracy. Despite the importance of high-resolution DEM, many areas in the United States and the world do not have access to high-resolution DEM due to technological limitations or the cost of the data collection. With recent development in Graphical Processing Units (GPU) and novel algorithms, deep learning techniques have become attractive to researchers for their performance in learning features from high-resolution datasets. Numerous new methods have been proposed such as Generative Adversarial Networks (GANs) to create intelligent models that correct and augment large-scale datasets. In this paper, a GAN based model is developed and evaluated, inspired by single image super-resolution methods, to increase the spatial resolution of a given DEM dataset up to 4 times without additional information related to data.
We introduce Geomstats, an open-source Python toolbox for computations and statistics on nonlinear manifolds, such as hyperbolic spaces, spaces of symmetric positive definite matrices, Lie groups of transformations, and many more. We provide object-oriented and extensively unit-tested implementations. Among others, manifolds come equipped with families of Riemannian metrics, with associated exponential and logarithmic maps, geodesics and parallel transport. Statistics and learning algorithms provide methods for estimation, clustering and dimension reduction on manifolds. All associated operations are vectorized for batch computation and provide support for different execution backends, namely NumPy, PyTorch and TensorFlow, enabling GPU acceleration. This paper presents the package, compares it with related libraries and provides relevant code examples. We show that Geomstats provides reliable building blocks to foster research in differential geometry and statistics, and to democratize the use of Riemannian geometry in machine learning applications. The source code is freely available under the MIT license at \url{geomstats.ai}.
In the contemporary big data realm, Deep Neural Networks (DNNs) are evolving towards more complex architectures to achieve higher inference accuracy. Model compression techniques can be leveraged to efficiently deploy such compute-intensive architectures on resource-limited mobile devices. Such methods comprise various hyper-parameters that require per-layer customization to ensure high accuracy. Choosing such hyper-parameters is cumbersome as the pertinent search space grows exponentially with model layers. This paper introduces GeneCAI, a novel optimization method that automatically learns how to tune per-layer compression hyper-parameters. We devise a bijective translation scheme that encodes compressed DNNs to the genotype space. The optimality of each genotype is measured using a multi-objective score based on accuracy and number of floating point operations. We develop customized genetic operations to iteratively evolve the non-dominated solutions towards the optimal Pareto front, thus, capturing the optimal trade-off between model accuracy and complexity. GeneCAI optimization method is highly scalable and can achieve a near-linear performance boost on distributed multi-GPU platforms. Our extensive evaluations demonstrate that GeneCAI outperforms existing rule-based and reinforcement learning methods in DNN compression by finding models that lie on a better accuracy-complexity Pareto curve.