PyTorchKR Papers

최신 논문

총 2,153편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:1810.06860공식 코드 · Papers with Code

Faster Matrix Completion Using Randomized SVD

Matrix completion is a widely used technique for image inpainting and personalized recommender system, etc. In this work, we focus on accelerating the matrix completion using faster randomized singular value decomposition (rSVD). Firstly, two fast randomized algorithms (rSVD-PI and rSVD- BKI) are proposed for handling sparse matrix. They make use of an eigSVD procedure and several accelerating skills. Then, with the rSVD-BKI algorithm and a new subspace recycling technique, we accelerate the singular value thresholding (SVT) method in [1] to realize faster matrix completion. Experiments show that the proposed rSVD algorithms can be 6X faster than the basic rSVD algorithm [2] while keeping same accuracy. For image inpainting and movie-rating estimation problems, the proposed accelerated SVT algorithm consumes 15X and 8X less CPU time than the methods using svds and lansvd respectively, without loss of accuracy.

arXiv:1810.05670

Computational ghost imaging using a field-programmable gate array

Computational ghost imaging is a promising technique for single-pixel imaging because it is robust to disturbance and can be operated over broad wavelength bands, unlike common cameras. However, one disadvantage of this method is that it has a long calculation time for image reconstruction. In this paper, we have designed a dedicated calculation circuit that accelerated the process of computational ghost imaging. We implemented this circuit by using a field-programmable gate array, which reduced the calculation time for the circuit compared to a CPU. The dedicated circuit reconstructs images at a frame rate of 300 Hz.

arXiv:1810.03988

GPU based Parallel Optimization for Real Time Panoramic Video Stitching

Panoramic video is a sort of video recorded at the same point of view to record the full scene. With the development of video surveillance and the requirement for 3D converged video surveillance in smart cities, CPU and GPU are required to possess strong processing abilities to make panoramic video. The traditional panoramic products depend on post processing, which results in high power consumption, low stability and unsatisfying performance in real time. In order to solve these problems,we propose a real-time panoramic video stitching framework.The framework we propose mainly consists of three algorithms, LORB image feature extraction algorithm, feature point matching algorithm based on LSH and GPU parallel video stitching algorithm based on CUDA.The experiment results show that the algorithm mentioned can improve the performance in the stages of feature extraction of images stitching and matching, the running speed of which is 11 times than that of the traditional ORB algorithm and 639 times than that of the traditional SIFT algorithm. Based on analyzing the GPU resources occupancy rate of each resolution image stitching, we further propose a stream parallel strategy to maximize the utilization of GPU resources. Compared with the L-ORB algorithm, the efficiency of this strategy is improved by 1.6-2.5 times, and it can make full use of GPU resources. The performance of the system accomplished in the paper is 29.2 times than that of the former embedded one, while the power dissipation is reduced to 10W.

arXiv:1810.04329

Task Runtime Prediction in Scientific Workflows Using an Online Incremental Learning Approach

Many algorithms in workflow scheduling and resource provisioning rely on the performance estimation of tasks to produce a scheduling plan. A profiler that is capable of modeling the execution of tasks and predicting their runtime accurately, therefore, becomes an essential part of any Workflow Management System (WMS). With the emergence of multi-tenant Workflow as a Service (WaaS) platforms that use clouds for deploying scientific workflows, task runtime prediction becomes more challenging because it requires the processing of a significant amount of data in a near real-time scenario while dealing with the performance variability of cloud resources. Hence, relying on methods such as profiling tasks' execution data using basic statistical description (e.g., mean, standard deviation) or batch offline regression techniques to estimate the runtime may not be suitable for such environments. In this paper, we propose an online incremental learning approach to predict the runtime of tasks in scientific workflows in clouds. To improve the performance of the predictions, we harness fine-grained resources monitoring data in the form of time-series records of CPU utilization, memory usage, and I/O activities that are reflecting the unique characteristics of a task's execution. We compare our solution to a state-of-the-art approach that exploits the resources monitoring data based on regression machine learning technique. From our experiments, the proposed strategy improves the performance, in terms of the error, up to 29.89%, compared to the state-of-the-art solutions.

arXiv:1810.02679

Distributed optimization in wireless sensor networks: an island-model framework

Wireless Sensor Networks (WSNs) is an emerging technology in several application domains, ranging from urban surveillance to environmental and structural monitoring. Computational Intelligence (CI) techniques are particularly suitable for enhancing these systems. However, when embedding CI into wireless sensors, severe hardware limitations must be taken into account. In this paper we investigate the possibility to perform an online, distributed optimization process within a WSN. Such a system might be used, for example, to implement advanced network features like distributed modelling, self-optimizing protocols, and anomaly detection, to name a few. The proposed approach, called DOWSN (Distributed Optimization for WSN) is an island-model infrastructure in which each node executes a simple, computationally cheap (both in terms of CPU and memory) optimization algorithm, and shares promising solutions with its neighbors. We perform extensive tests of different DOWSN configurations on a benchmark made up of continuous optimization problems; we analyze the influence of the network parameters (number of nodes, inter-node communication period and probability of accepting incoming solutions) on the optimization performance. Finally, we profile energy and memory consumption of DOWSN to show the efficient usage of the limited hardware resources available on the sensor nodes.

arXiv:1810.02648

LiveCap: Real-time Human Performance Capture from Monocular Video

We present the first real-time human performance capture approach that reconstructs dense, space-time coherent deforming geometry of entire humans in general everyday clothing from just a single RGB video. We propose a novel two-stage analysis-by-synthesis optimization whose formulation and implementation are designed for high performance. In the first stage, a skinned template model is jointly fitted to background subtracted input video, 2D and 3D skeleton joint positions found using a deep neural network, and a set of sparse facial landmark detections. In the second stage, dense non-rigid 3D deformations of skin and even loose apparel are captured based on a novel real-time capable algorithm for non-rigid tracking using dense photometric and silhouette constraints. Our novel energy formulation leverages automatically identified material regions on the template to model the differing non-rigid deformation behavior of skin and apparel. The two resulting non-linear optimization problems per-frame are solved with specially-tailored data-parallel Gauss-Newton solvers. In order to achieve real-time performance of over 25Hz, we design a pipelined parallel architecture using the CPU and two commodity GPUs. Our method is the first real-time monocular approach for full-body performance capture. Our method yields comparable accuracy with off-line performance capture techniques, while being orders of magnitude faster.

arXiv:1809.09095

On Reinforcement Learning for Full-length Game of StarCraft

StarCraft II poses a grand challenge for reinforcement learning. The main difficulties of it include huge state and action space and a long-time horizon. In this paper, we investigate a hierarchical reinforcement learning approach for StarCraft II. The hierarchy involves two levels of abstraction. One is the macro-action automatically extracted from expert's trajectories, which reduces the action space in an order of magnitude yet remains effective. The other is a two-layer hierarchical architecture which is modular and easy to scale, enabling a curriculum transferring from simpler tasks to more complex tasks. The reinforcement training algorithm for this architecture is also investigated. On a 64x64 map and using restrictive units, we achieve a winning rate of more than 99\% against the difficulty level-1 built-in AI. Through the curriculum transfer learning algorithm and a mixture of combat model, we can achieve over 93\% winning rate of Protoss against the most difficult non-cheating built-in AI (level-7) of Terran, training within two days using a single machine with only 48 CPU cores and 8 K40 GPUs. It also shows strong generalization performance, when tested against never seen opponents including cheating levels built-in AI and all levels of Zerg and Protoss built-in AI. We hope this study could shed some light on the future research of large-scale reinforcement learning.

arXiv:1809.07406

Exploiting Tournament Selection for Efficient Parallel Genetic Programming

Genetic Programming (GP) is a computationally intensive technique which is naturally parallel in nature. Consequently, many attempts have been made to improve its run-time from exploiting highly parallel hardware such as GPUs. However, a second methodology of improving the speed of GP is through efficiency techniques such as subtree caching. However achieving parallel performance and efficiency is a difficult task. This paper will demonstrate an efficiency saving for GP compatible with the harnessing of parallel CPU hardware by exploiting tournament selection. Significant efficiency savings are demonstrated whilst retaining the capability of a high performance parallel implementation of GP. Indeed, a 74% improvement in the speed of GP is achieved with a peak rate of 96 billion GPop/s for classification type problems.

arXiv:1809.07196공식 코드 · Papers with Code

Characterising Across-Stack Optimisations for Deep Convolutional Neural Networks

Convolutional Neural Networks (CNNs) are extremely computationally demanding, presenting a large barrier to their deployment on resource-constrained devices. Since such systems are where some of their most useful applications lie (e.g. obstacle detection for mobile robots, vision-based medical assistive technology), significant bodies of work from both machine learning and systems communities have attempted to provide optimisations that will make CNNs available to edge devices. In this paper we unify the two viewpoints in a Deep Learning Inference Stack and take an across-stack approach by implementing and evaluating the most common neural network compression techniques (weight pruning, channel pruning, and quantisation) and optimising their parallel execution with a range of programming approaches (OpenMP, OpenCL) and hardware architectures (CPU, GPU). We provide comprehensive Pareto curves to instruct trade-offs under constraints of accuracy, execution time, and memory space.

arXiv:1809.03668

Comparing Computing Platforms for Deep Learning on a Humanoid Robot

The goal of this study is to test two different computing platforms with respect to their suitability for running deep networks as part of a humanoid robot software system. One of the platforms is the CPU-centered Intel NUC7i7BNH and the other is a NVIDIA Jetson TX2 system that puts more emphasis on GPU processing. The experiments addressed a number of benchmarking tasks including pedestrian detection using deep neural networks. Some of the results were unexpected but demonstrate that platforms exhibit both advantages and disadvantages when taking computational performance and electrical power requirements of such a system into account.

arXiv:1609.06634

BioEM: GPU-accelerated computing of Bayesian inference of electron microscopy images

In cryo-electron microscopy (EM), molecular structures are determined from large numbers of projection images of individual particles. To harness the full power of this single-molecule information, we use the Bayesian inference of EM (BioEM) formalism. By ranking structural models using posterior probabilities calculated for individual images, BioEM in principle addresses the challenge of working with highly dynamic or heterogeneous systems not easily handled in traditional EM reconstruction. However, the calculation of these posteriors for large numbers of particles and models is computationally demanding. Here we present highly parallelized, GPU-accelerated computer software that performs this task efficiently. Our flexible formulation employs CUDA, OpenMP, and MPI parallelization combined with both CPU and GPU computing. The resulting BioEM software scales nearly ideally both on pure CPU and on CPU+GPU architectures, thus enabling Bayesian analysis of tens of thousands of images in a reasonable time. The general mathematical framework and robust algorithms are not limited to cryo-electron microscopy but can be generalized for electron tomography and other imaging experiments.

arXiv:1808.03843공식 코드 · Papers with Code

Matrix Factorization on GPUs with Memory Optimization and Approximate Computing

Matrix factorization (MF) discovers latent features from observations, which has shown great promises in the fields of collaborative filtering, data compression, feature extraction, word embedding, etc. While many problem-specific optimization techniques have been proposed, alternating least square (ALS) remains popular due to its general applicability e.g. easy to handle positive-unlabeled inputs, fast convergence and parallelization capability. Current MF implementations are either optimized for a single machine or with a need of a large computer cluster but still are insufficient. This is because a single machine provides limited compute power for large-scale data while multiple machines suffer from the network communication bottleneck. To address the aforementioned challenge, accelerating ALS on graphics processing units (GPUs) is a promising direction. We propose the novel approach in enhancing the MF efficiency via both memory optimization and approximate computing. The former exploits GPU memory hierarchy to increase data reuse, while the later reduces unnecessary computing without hurting the convergence of learning algorithms. Extensive experiments on large-scale datasets show that our solution not only outperforms the competing CPU solutions by a large margin but also has a 2x-4x performance gain compared to the state-of-the-art GPU solutions. Our implementations are open-sourced and publicly available.

arXiv:1808.03001

Compressed Sensing Using Binary Matrices of Nearly Optimal Dimensions

In this paper, we study the problem of compressed sensing using binary measurement matrices and $\ell_1$-norm minimization (basis pursuit) as the recovery algorithm. We derive new upper and lower bounds on the number of measurements to achieve robust sparse recovery with binary matrices. We establish sufficient conditions for a column-regular binary matrix to satisfy the robust null space property (RNSP) and show that the associated sufficient conditions % sparsity bounds for robust sparse recovery obtained using the RNSP are better by a factor of $(3 \sqrt{3})/2 \approx 2.6$ compared to the sufficient conditions obtained using the restricted isometry property (RIP). Next we derive universal \textit{lower} bounds on the number of measurements that any binary matrix needs to have in order to satisfy the weaker sufficient condition based on the RNSP and show that bipartite graphs of girth six are optimal. Then we display two classes of binary matrices, namely parity check matrices of array codes and Euler squares, which have girth six and are nearly optimal in the sense of almost satisfying the lower bound. In principle, randomly generated Gaussian measurement matrices are "order-optimal". So we compare the phase transition behavior of the basis pursuit formulation using binary array codes and Gaussian matrices and show that (i) there is essentially no difference between the phase transition boundaries in the two cases and (ii) the CPU time of basis pursuit with binary matrices is hundreds of times faster than with Gaussian matrices and the storage requirements are less. Therefore it is suggested that binary matrices are a viable alternative to Gaussian matrices for compressed sensing using basis pursuit. \end{abstract}

arXiv:1808.01363

GeneSys: Enabling Continuous Learning through Neural Network Evolution in Hardware

Modern deep learning systems rely on (a) a hand-tuned neural network topology, (b) massive amounts of labeled training data, and (c) extensive training over large-scale compute resources to build a system that can perform efficient image classification or speech recognition. Unfortunately, we are still far away from implementing adaptive general purpose intelligent systems which would need to learn autonomously in unknown environments and may not have access to some or any of these three components. Reinforcement learning and evolutionary algorithm (EA) based methods circumvent this problem by continuously interacting with the environment and updating the models based on obtained rewards. However, deploying these algorithms on ubiquitous autonomous agents at the edge (robots/drones) demands extremely high energy-efficiency due to (i) tight power and energy budgets, (ii) continuous/lifelong interaction with the environment, (iii) intermittent or no connectivity to the cloud to run heavy-weight processing. To address this need, we present GENESYS, an HW-SW prototype of an EA-based learning system, that comprises a closed loop learning engine called EvE and an inference engine called ADAM. EvE can evolve the topology and weights of neural networks completely in hardware for the task at hand, without requiring hand-optimization or backpropagation training. ADAM continuously interacts with the environment and is optimized for efficiently running the irregular neural networks generated by EvE. GENESYS identifies and leverages multiple unique avenues of parallelism unique to EAs that we term 'gene'- level parallelism, and 'population'-level parallelism. We ran GENESYS with a suite of environments from OpenAI gym and observed 2-5 orders of magnitude higher energy-efficiency over state-of-the-art embedded and desktop CPU and GPU systems.

arXiv:1807.11091공식 코드 · Papers with Code

StructADMM: A Systematic, High-Efficiency Framework of Structured Weight Pruning for DNNs

Weight pruning methods of DNNs have been demonstrated to achieve a good model pruning rate without loss of accuracy, thereby alleviating the significant computation/storage requirements of large-scale DNNs. Structured weight pruning methods have been proposed to overcome the limitation of irregular network structure and demonstrated actual GPU acceleration. However, in prior work the pruning rate (degree of sparsity) and GPU acceleration are limited (to less than 50%) when accuracy needs to be maintained. In this work,we overcome these limitations by proposing a unified, systematic framework of structured weight pruning for DNNs. It is a framework that can be used to induce different types of structured sparsity, such as filter-wise, channel-wise, and shape-wise sparsity, as well non-structured sparsity. The proposed framework incorporates stochastic gradient descent with ADMM, and can be understood as a dynamic regularization method in which the regularization target is analytically updated in each iteration. Without loss of accuracy on the AlexNet model, we achieve 2.58X and 3.65X average measured speedup on two GPUs, clearly outperforming the prior work. The average speedups reach 3.15X and 8.52X when allowing a moderate ac-curacy loss of 2%. In this case the model compression for convolutional layers is 15.0X, corresponding to 11.93X measured CPU speedup. Our experiments on ResNet model and on other data sets like UCF101 and CIFAR-10 demonstrate the consistently higher performance of our framework.

arXiv:1807.08271공식 코드 · Papers with Code

RGBiD-SLAM for Accurate Real-time Localisation and 3D Mapping

In this paper we present a complete SLAM system for RGB-D cameras, namely RGB-iD SLAM. The presented approach is a dense direct SLAM method with the main characteristic of working with the depth maps in inverse depth parametrisation for the routines of dense alignment or keyframe fusion. The system consists in 2 CPU threads working in parallel, which share the use of the GPU for dense alignment and keyframe fusion routines. The first thread is a front-end operating at frame rate, which processes every incoming frame from the RGB-D sensor to compute the incremental odometry and integrate it in a keyframe which is changed periodically following a covisibility-based strategy. The second thread is a back-end which receives keyframes from the front-end. This thread is in charge of segmenting the keyframes based on their structure, describing them using Bags of Words, trying to find potential loop closures with previous keyframes, and in such case perform pose-graph optimisation for trajectory correction. In addition, our system allows is able to compute the odometry both with unregistered and registered depth maps, allowing to use customised calibrations of the RGB-D sensor. As a consequence in the paper we also propose a detailed calibration pipeline to compute customised calibrations for particular RGB-D cameras. The experiments with our approach in the TUM RGB-D benchmark datasets show results superior in accuracy to the state-of-the-art in many of the sequences. The code has been made available on-line for research purposes https://github.com/dangut/RGBiD-SLAM.

arXiv:1807.06555

Training Recurrent Neural Networks against Noisy Computations during Inference

We explore the robustness of recurrent neural networks when the computations within the network are noisy. One of the motivations for looking into this problem is to reduce the high power cost of conventional computing of neural network operations through the use of analog neuromorphic circuits. Traditional GPU/CPU-centered deep learning architectures exhibit bottlenecks in power-restricted applications, such as speech recognition in embedded systems. The use of specialized neuromorphic circuits, where analog signals passed through memory-cell arrays are sensed to accomplish matrix-vector multiplications, promises large power savings and speed gains but brings with it the problems of limited precision of computations and unavoidable analog noise. In this paper we propose a method, called {\em Deep Noise Injection training}, to train RNNs to obtain a set of weights/biases that is much more robust against noisy computation during inference. We explore several RNN architectures, such as vanilla RNN and long-short-term memories (LSTM), and show that after convergence of Deep Noise Injection training the set of trained weights/biases has more consistent performance over a wide range of noise powers entering the network during inference. Surprisingly, we find that Deep Noise Injection training improves overall performance of some networks even for numerically accurate inference.

arXiv:1807.02752

Real-time stereo vision-based lane detection system

The detection of multiple curved lane markings on a non-flat road surface is still a challenging task for automotive applications. To make an improvement, the depth information can be used to greatly enhance the robustness of the lane detection systems. The proposed system in this paper is developed from our previous work where the dense vanishing point Vp is estimated globally to assist the detection of multiple curved lane markings. However, the outliers in the optimal solution may severely affect the accuracy of the least squares fitting when estimating Vp. Therefore, in this paper we use Random Sample Consensus to update the inliers and outliers iteratively until the fraction of the number of inliers versus the total number exceeds our pre-set threshold. This significantly helps the system to overcome some suddenly changing conditions. Furthermore, we propose a novel lane position validation approach which provides a piecewise weight based on Vp and the gradient to reduce the gradient magnitude of the non-lane candidates. Then, we compute the energy of each possible solution and select all satisfying lane positions for visualisation. The proposed system is implemented on a heterogeneous system which consists of an Intel Core i7-4720HQ CPU and a NVIDIA GTX 970M GPU. A processing speed of 143 fps has been achieved, which is over 38 times faster than our previous work. Also, in order to evaluate the detection precision, we tested 2495 frames with 5361 lanes from the KITTI database (1637 lanes more than our previous experiment). It is shown that the overall successful detection rate is improved from 98.7% to 99.5%.

arXiv:1807.02037

TFLMS: Large Model Support in TensorFlow by Graph Rewriting

While accelerators such as GPUs have limited memory, deep neural networks are becoming larger and will not fit with the memory limitation of accelerators for training. We propose an approach to tackle this problem by rewriting the computational graph of a neural network, in which swap-out and swap-in operations are inserted to temporarily store intermediate results on CPU memory. In particular, we first revise the concept of a computational graph by defining a concrete semantics for variables in a graph. We then formally show how to derive swap-out and swap-in operations from an existing graph and present rules to optimize the graph. To realize our approach, we developed a module in TensorFlow, named TFLMS. TFLMS is published as a pull request in the TensorFlow repository for contributing to the TensorFlow community. With TFLMS, we were able to train ResNet-50 and 3DUnet with 4.7x and 2x larger batch size, respectively. In particular, we were able to train 3DUNet using images of size of $192^3$ for image segmentation, which, without TFLMS, had been done only by dividing the images to smaller images, which affects the accuracy.

arXiv:1807.01350

OCTen: Online Compression-based Tensor Decomposition

Tensor decompositions are powerful tools for large data analytics as they jointly model multiple aspects of data into one framework and enable the discovery of the latent structures and higher-order correlations within the data. One of the most widely studied and used decompositions, especially in data mining and machine learning, is the Canonical Polyadic or CP decomposition. However, today's datasets are not static and these datasets often dynamically growing and changing with time. To operate on such large data, we present OCTen the first ever compression-based online parallel implementation for the CP decomposition. We conduct an extensive empirical analysis of the algorithms in terms of fitness, memory used and CPU time, and in order to demonstrate the compression and scalability of the method, we apply OCTen to big tensor data. Indicatively, OCTen performs on-par or better than state-of-the-art online and online methods in terms of decomposition accuracy and efficiency, while saving up to 40-200 % memory space.