DOCSLIB.ORG

VO2 Carbon Nanotube Composite Memristor-Based Cellular Neural Network Pattern Formation

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
VO2 Carbon Nanotube Composite Memristor-Based Cellular Neural Network Pattern Formation electronics Article VO2 Carbon Nanotube Composite Memristor-Based Cellular Neural Network Pattern Formation Yiran Shen and Guangyi Wang * Institute of Modern Circuits and Intelligent Information, Hangzhou Dianzi University, Hangzhou 310018, China; [email protected] * Correspondence: [email protected] Abstract: A cellular neural network (CNN) based on a VO2 carbon nanotube memristor is proposed in this paper. The device is modeled by SPICE at first, and then the cell dynamic characteristics based on the device are analyzed. It is pointed out that only when the cell is at the sharp edge of chaos can the cell be successfully awakened after the CNN is formed. In this paper, we give the example of a 5 × 5 CNN, set specific initial conditions and observe the formed pattern. Because the generated patterns are affected by the initial conditions, the cell power supply can be pre-programmed to obtain specific patterns, which can be applied to the future information processing system based on complex space–time patterns, especially in the field of computer vision. Keywords: VO2 carbon nanotube composite memristor; cellular neural network (CNN); von Neu- mann structure; local activity; edge of chaos 1. Introduction Citation: Shen, Y.; Wang, G. VO2 The traditional processor uses the von Neumann structure, in which the storage unit Carbon Nanotube Composite and the processing unit are separated and connected by bus. In recent years, with the Memristor-Based Cellular Neural development of semiconductor technology, the speed of processing units has been greatly Network Pattern Formation. improved. However, due to the limitation of bus bandwidth, the operation speed of the Electronics 2021, 10, 1198. https:// whole processor is limited, which is called the “von Neumann bottleneck problem” [1,2]. In doi.org/10.3390/electronics10101198 order to solve this problem, inspired by the biological way of processing information, many pieces of literature have proposed various non-von Neumann processor solutions [3–5]. Academic Editor: Kris Campbell Typical bio-inspired computing relies on nonlinear networks that contain the same cells; each cell has a relatively simple structure and interacts with the surrounding cells. This Received: 10 April 2021 structure, called a CNN, a cellular neural network, has been applied in fields such as Accepted: 11 May 2021 Published: 18 May 2021 computer vision [6–8]. The concept of a CNN can be traced back to two articles by Chua in 1988 [9,10], which Publisher’s Note: MDPI stays neutral respectively give the theory and application of CNNs. Recently, Itoh [11] summarized with regard to jurisdictional claims in some characteristics of CNNs in a long paper. Weiher et al. [12] discussed the pattern published maps and institutional affil- formation of CNNs based on the NbO2 memristor. iations. Recently, brain-like computing has become a hot topic in research [13–15], and the key is to find devices that can produce spike signals like neurons and that have very low power consumption. The VO2 carbon nanotube (CNT) composite device is a Mott memristor recently proposed [16]. It can generate periodic peak pulses with a pulse width of less than 20 ns, it is driven by a DC current or voltage, and it does not need additional Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. capacitance. It uses metal–carbon nanotubes as heaters, and, compared with the pure Mott This article is an open access article VO2 proposed earlier, adding CNTs can greatly reduce the transient duration and pulse distributed under the terms and energy and increase the frequency of a peak pulse by three orders of magnitude. conditions of the Creative Commons The VO2 nano crossbar device does not need the process of electric forming and has Attribution (CC BY) license (https:// low device size dispersion. For devices with a critical size between 50 and 600 nm, the creativecommons.org/licenses/by/ change coefficient of the switch threshold voltage is less than 13%, the switch durability 4.0/). is more than 26.6 million cycles, and the IV characteristics of the device are not changed Electronics 2021, 10, 1198. https://doi.org/10.3390/electronics10101198 https://www.mdpi.com/journal/electronics Electronics 2021, 10, x FOR PEER REVIEW 2 of 14 Electronics 2021, 10, 1198 2 of 13 change coefficient of the switch threshold voltage is less than 13%, the switch durability is more than 26.6 million cycles, and the IV characteristics of the device are not changed significantly. The VO2 device technology in the process of non-electric formation acceler- significantly.ates the development The VO2 device of an technologyactive memristor in the process neuron of circuit. non-electric It ca formationn simulate accelerates the most theknown development neuron dynamics of an active and memristorclear the way neuron for the circuit. realization It can simulateof a large-scale the most integrated known neuron dynamics and clear the way for the realization of a large-scale integrated circuit circuit (IC). In addition, the VO2 memristor is superior to its NbO2 counterpart in both (IC). In addition, the VO2 memristor is superior to its NbO2 counterpart in both switch switch speed and switching energy. The simulated Mott transition in the VO2 is 100 times speed and switching energy. The simulated Mott transition in the VO2 is 100 times faster faster than in the NbO2 and consumes only about one-sixth (16%) of the energy [17]. than in the NbO2 and consumes only about one-sixth (16%) of the energy [17]. In this paper, a cellular neural network based on a VO2 CNT is proposed. As a cell In this paper, a cellular neural network based on a VO2 CNT is proposed. As a cell itself, a VO2 CNT is set to be stable and static, that is, in “sleep”. If the memristor is on the itself, a VO CNT is set to be stable and static, that is, in “sleep”. If the memristor is “edge of chaos”,2 two or more cells are connected by the RC coupling, which will make it on the “edge of chaos”, two or more cells are connected by the RC coupling, which will in the state of “wake-up”, namely, in the dynamic oscillation mode. The second part of make it in the state of “wake-up”, namely, in the dynamic oscillation mode. The second this paper is the modeling of VO2 CNT composite devices, providing the SPICE model. part of this paper is the modeling of VO CNT composite devices, providing the SPICE The third part is the cell circuit, which gives2 the analysis process of the decoupled circuit. model. The third part is the cell circuit, which gives the analysis process of the decoupled circuit.The response The response of the memristor of the memristor is expanded is expanded near the near operating the operating point and point the and small-signal the small- signalequivalent equivalent circuit circuitis given. is Based given. on Based this, on the this, influence the influence of coupling of coupling R and CR deviceand C param-device parameterseters on the oninput the impedance input impedance of one port of one is analyzed. port is analyzed. The fourth The part fourth is the part CNN is the simula- CNN simulation,tion, which describes which describes the pattern the patternformation formation characteristics characteristics of a CNN of composed a CNN composed of memris- of memristor-basedtor-based cells. cells. Compared with the traditional CMOS realization of CNN, the memristor counterpart can largely save power and chipchip area,area, andand provideprovide ultra-highultra-high processingprocessing speeds.speeds. 2. VO 2 CarbonCarbon Nanotube Nanotube Compos Compositeite Device Device Modeling The structure of of the the VO VO22 carboncarbon nanotube nanotube composite composite device device is isshown shown in inFigure Figure 1.1 It. Itcontains contains a transverse a transverse active active region, region, which which is defined is defined by a byVO a2 VOthin2 metalthin metal strip with strip a with thick- a thicknessness of about of about 5 nm 5(as nm shown (as shown in the in red the region red region of the of figure), the figure), and its and two its ends two endsare con- are connectednected with with Pd electrodes Pd electrodes (as shown (as shown in the in blue the blueregion region of the of figure). the figure). This is This a planar is a planar Mott Mottmetal-insulator metal-insulator transition transition device. device. The aligned The aligned carbon carbon nanotubes nanotubes (as shown (as shown in the inblack the blackline) were line) first were grown first grown on quartz on quartz substrate, substrate, and then and thentransferred transferred to the to surface the surface of a VO of a2 VOthin2 metalthin metal strip stripbefore before the whole the whole device device was formed. was formed. Figure 1. A VO 2 carboncarbon nanotube composite device structure. The redred partpart isis thethe VOVO22 and the black lineblack over line it over represents it represents the carbon the carbon nanotube. nanotube. As a Mott metal-insulator transition device, it satisfiessatisfies thethe followingfollowing equationequation [[16]:16]: p βφvdb− vm/d−f 2 m( ) i = ATekT kT iATe=m 2 m i v +v2 /R T−T (1) dT = m m m CNT − amb (1) dt ivC+−th v2 RC Tth R Tth dT =−mm m CNT amb dtC CR The first formula of Equation (1) is the currentth emission th th equation of the Schottky diode, and the second formula is Newton’s cooling law, where, im and νm are the current and voltageThe through first formula the memristor of Equation NDR (1) device, is the cu respectively;rrent emissionT is equation the absolute of the temperature Schottky di- of theode, device; and theT ambsecondis the formula ambient is temperature;Newton's coolingA and law,b are where, scaling ݅m constants;and ݒm ared theis thecurrent effective and devicevoltage length;throughk theis Boltzmann memristor constant;NDR device,f is respectively; the energy barrier;T is the absoluteCth is the temperature heat capacity of (effectivethe device; thermal Tamb is the mass); ambientRth is temperature; the thermal A resistance; and ߚ areR scalingCNT is theconstants; resistance d is valuethe effective of the CNT (carbon nanotube), taking 600 kW.
Load more

Recommended publications
  • Synthesizing Images of Humans in Unseen Poses
    Synthesizing Images of Humans in Unseen Poses Guha Balakrishnan Amy Zhao Adrian V. Dalca Fredo Durand MIT MIT MIT and MGH MIT [email protected] [email protected] [email protected] [email protected] John Guttag MIT [email protected] Abstract We address the computational problem of novel human pose synthesis. Given an image of a person and a desired pose, we produce a depiction of that person in that pose, re- taining the appearance of both the person and background. We present a modular generative neural network that syn- Source Image Target Pose Synthesized Image thesizes unseen poses using training pairs of images and poses taken from human action videos. Our network sepa- Figure 1. Our method takes an input image along with a desired target pose, and automatically synthesizes a new image depicting rates a scene into different body part and background lay- the person in that pose. We retain the person’s appearance as well ers, moves body parts to new locations and refines their as filling in appropriate background textures. appearances, and composites the new foreground with a hole-filled background. These subtasks, implemented with separate modules, are trained jointly using only a single part details consistent with the new pose. Differences in target image as a supervised label. We use an adversarial poses can cause complex changes in the image space, in- discriminator to force our network to synthesize realistic volving several moving parts and self-occlusions. Subtle details conditioned on pose. We demonstrate image syn- details such as shading and edges should perceptually agree thesis results on three action classes: golf, yoga/workouts with the body’s configuration.
    [Show full text]
  • Memristor-Based Approximated Computation
    Memristor-based Approximated Computation Boxun Li1, Yi Shan1, Miao Hu2, Yu Wang1, Yiran Chen2, Huazhong Yang1 1Dept. of E.E., TNList, Tsinghua University, Beijing, China 2Dept. of E.C.E., University of Pittsburgh, Pittsburgh, USA 1 Email: [email protected] Abstract—The cessation of Moore’s Law has limited further architectures, which not only provide a promising hardware improvements in power efficiency. In recent years, the physical solution to neuromorphic system but also help drastically close realization of the memristor has demonstrated a promising the gap of power efficiency between computing systems and solution to ultra-integrated hardware realization of neural net- works, which can be leveraged for better performance and the brain. The memristor is one of those promising devices. power efficiency gains. In this work, we introduce a power The memristor is able to support a large number of signal efficient framework for approximated computations by taking connections within a small footprint by taking the advantage advantage of the memristor-based multilayer neural networks. of the ultra-integration density [7]. And most importantly, A programmable memristor approximated computation unit the nonvolatile feature that the state of the memristor could (Memristor ACU) is introduced first to accelerate approximated computation and a memristor-based approximated computation be tuned by the current passing through itself makes the framework with scalability is proposed on top of the Memristor memristor a potential, perhaps even the best, device to realize ACU. We also introduce a parameter configuration algorithm of neuromorphic computing systems with picojoule level energy the Memristor ACU and a feedback state tuning circuit to pro- consumption [8], [9].
    [Show full text]
  • A Survey of Autonomous Driving: Common Practices and Emerging Technologies
    Accepted March 22, 2020 Digital Object Identifier 10.1109/ACCESS.2020.2983149 A Survey of Autonomous Driving: Common Practices and Emerging Technologies EKIM YURTSEVER1, (Member, IEEE), JACOB LAMBERT 1, ALEXANDER CARBALLO 1, (Member, IEEE), AND KAZUYA TAKEDA 1, 2, (Senior Member, IEEE) 1Nagoya University, Furo-cho, Nagoya, 464-8603, Japan 2Tier4 Inc. Nagoya, Japan Corresponding author: Ekim Yurtsever (e-mail: [email protected]). ABSTRACT Automated driving systems (ADSs) promise a safe, comfortable and efficient driving experience. However, fatalities involving vehicles equipped with ADSs are on the rise. The full potential of ADSs cannot be realized unless the robustness of state-of-the-art is improved further. This paper discusses unsolved problems and surveys the technical aspect of automated driving. Studies regarding present challenges, high- level system architectures, emerging methodologies and core functions including localization, mapping, perception, planning, and human machine interfaces, were thoroughly reviewed. Furthermore, many state- of-the-art algorithms were implemented and compared on our own platform in a real-world driving setting. The paper concludes with an overview of available datasets and tools for ADS development. INDEX TERMS Autonomous Vehicles, Control, Robotics, Automation, Intelligent Vehicles, Intelligent Transportation Systems I. INTRODUCTION necessary here. CCORDING to a recent technical report by the Eureka Project PROMETHEUS [11] was carried out in A National Highway Traffic Safety Administration Europe between 1987-1995, and it was one of the earliest (NHTSA), 94% of road accidents are caused by human major automated driving studies. The project led to the errors [1]. Against this backdrop, Automated Driving Sys- development of VITA II by Daimler-Benz, which succeeded tems (ADSs) are being developed with the promise of in automatically driving on highways [12].
    [Show full text]
  • CSE 152: Computer Vision Manmohan Chandraker
    CSE 152: Computer Vision Manmohan Chandraker Lecture 15: Optimization in CNNs Recap Engineered against learned features Label Convolutional filters are trained in a Dense supervised manner by back-propagating classification error Dense Dense Convolution + pool Label Convolution + pool Classifier Convolution + pool Pooling Convolution + pool Feature extraction Convolution + pool Image Image Jia-Bin Huang and Derek Hoiem, UIUC Two-layer perceptron network Slide credit: Pieter Abeel and Dan Klein Neural networks Non-linearity Activation functions Multi-layer neural network From fully connected to convolutional networks next layer image Convolutional layer Slide: Lazebnik Spatial filtering is convolution Convolutional Neural Networks [Slides credit: Efstratios Gavves] 2D spatial filters Filters over the whole image Weight sharing Insight: Images have similar features at various spatial locations! Key operations in a CNN Feature maps Spatial pooling Non-linearity Convolution (Learned) . Input Image Input Feature Map Source: R. Fergus, Y. LeCun Slide: Lazebnik Convolution as a feature extractor Key operations in a CNN Feature maps Rectified Linear Unit (ReLU) Spatial pooling Non-linearity Convolution (Learned) Input Image Source: R. Fergus, Y. LeCun Slide: Lazebnik Key operations in a CNN Feature maps Spatial pooling Max Non-linearity Convolution (Learned) Input Image Source: R. Fergus, Y. LeCun Slide: Lazebnik Pooling operations • Aggregate multiple values into a single value • Invariance to small transformations • Keep only most important information for next layer • Reduces the size of the next layer • Fewer parameters, faster computations • Observe larger receptive field in next layer • Hierarchically extract more abstract features Key operations in a CNN Feature maps Spatial pooling Non-linearity Convolution (Learned) . Input Image Input Feature Map Source: R.
    [Show full text]
  • 1 Convolution
    CS1114 Section 6: Convolution February 27th, 2013 1 Convolution Convolution is an important operation in signal and image processing. Convolution op- erates on two signals (in 1D) or two images (in 2D): you can think of one as the \input" signal (or image), and the other (called the kernel) as a “filter” on the input image, pro- ducing an output image (so convolution takes two images as input and produces a third as output). Convolution is an incredibly important concept in many areas of math and engineering (including computer vision, as we'll see later). Definition. Let's start with 1D convolution (a 1D \image," is also known as a signal, and can be represented by a regular 1D vector in Matlab). Let's call our input vector f and our kernel g, and say that f has length n, and g has length m. The convolution f ∗ g of f and g is defined as: m X (f ∗ g)(i) = g(j) · f(i − j + m=2) j=1 One way to think of this operation is that we're sliding the kernel over the input image. For each position of the kernel, we multiply the overlapping values of the kernel and image together, and add up the results. This sum of products will be the value of the output image at the point in the input image where the kernel is centered. Let's look at a simple example. Suppose our input 1D image is: f = 10 50 60 10 20 40 30 and our kernel is: g = 1=3 1=3 1=3 Let's call the output image h.
    [Show full text]
  • Self-Driving Cars: Evaluation of Deep Learning Techniques for Object Detection in Different Driving Conditions Ramesh Simhambhatla SMU, [email protected]
    SMU Data Science Review Volume 2 | Number 1 Article 23 2019 Self-Driving Cars: Evaluation of Deep Learning Techniques for Object Detection in Different Driving Conditions Ramesh Simhambhatla SMU, [email protected] Kevin Okiah SMU, [email protected] Shravan Kuchkula SMU, [email protected] Robert Slater Southern Methodist University, [email protected] Follow this and additional works at: https://scholar.smu.edu/datasciencereview Part of the Artificial Intelligence and Robotics Commons, Other Computer Engineering Commons, and the Statistical Models Commons Recommended Citation Simhambhatla, Ramesh; Okiah, Kevin; Kuchkula, Shravan; and Slater, Robert (2019) "Self-Driving Cars: Evaluation of Deep Learning Techniques for Object Detection in Different Driving Conditions," SMU Data Science Review: Vol. 2 : No. 1 , Article 23. Available at: https://scholar.smu.edu/datasciencereview/vol2/iss1/23 This Article is brought to you for free and open access by SMU Scholar. It has been accepted for inclusion in SMU Data Science Review by an authorized administrator of SMU Scholar. For more information, please visit http://digitalrepository.smu.edu. Simhambhatla et al.: Self-Driving Cars: Evaluation of Deep Learning Techniques for Object Detection in Different Driving Conditions Self-Driving Cars: Evaluation of Deep Learning Techniques for Object Detection in Different Driving Conditions Kevin Okiah1, Shravan Kuchkula1, Ramesh Simhambhatla1, Robert Slater1 1 Master of Science in Data Science, Southern Methodist University, Dallas, TX 75275 USA {KOkiah, SKuchkula, RSimhambhatla, RSlater}@smu.edu Abstract. Deep Learning has revolutionized Computer Vision, and it is the core technology behind capabilities of a self-driving car. Convolutional Neural Networks (CNNs) are at the heart of this deep learning revolution for improving the task of object detection.
    [Show full text]
  • Do Better Imagenet Models Transfer Better?
    Do Better ImageNet Models Transfer Better? Simon Kornblith,∗ Jonathon Shlens, and Quoc V. Le Google Brain {skornblith,shlens,qvl}@google.com Logistic Regression Fine-Tuned Abstract 1.6 2.3 Inception-ResNet v2 Inception v4 1.5 2.2 Transfer learning is a cornerstone of computer vision, yet little work has been done to evaluate the relationship 1.4 MobileNet v1 2.1 MobileNet v1 NASNet Large NASNet Large between architecture and transfer. An implicit hypothesis 1.3 2.0 ResNet-50 ResNet-50 in modern computer vision research is that models that per- 1.2 1.9 form better on ImageNet necessarily perform better on other vision tasks. However, this hypothesis has never been sys- 1.1 1.8 Transfer Accuracy (Log Odds) tematically tested. Here, we compare the performance of 16 72 74 76 78 80 72 74 76 78 80 classification networks on 12 image classification datasets. ImageNet Top-1 Accuracy (%) ImageNet Top-1 Accuracy (%) We find that, when networks are used as fixed feature ex- Figure 1. Transfer learning performance is highly correlated with tractors or fine-tuned, there is a strong correlation between ImageNet top-1 accuracy for fixed ImageNet features (left) and ImageNet accuracy and transfer accuracy (r = 0.99 and fine-tuning from ImageNet initialization (right). The 16 points in 0.96, respectively). In the former setting, we find that this re- each plot represent transfer accuracy for 16 distinct CNN architec- tures, averaged across 12 datasets after logit transformation (see lationship is very sensitive to the way in which networks are Section 3).
    [Show full text]
  • Deep Learning for Computer Vision
    Deep Learning for Computer Vision MIT 6.S191 Ava Soleimany January 29, 2019 6.S191 Introduction to Deep Learning 1/29/19 introtodeeplearning.com What Computers “See” Images are Numbers 6.S191 Introduction to Deep Learning [1] 1/29/19 introtodeeplearning.com Images are Numbers 6.S191 Introduction to Deep Learning [1] 1/29/19 introtodeeplearning.com Images are Numbers What the computer sees An image is just a matrix of numbers [0,255]! i.e., 1080x1080x3 for an RGB image 6.S191 Introduction to Deep Learning [1] 1/29/19 introtodeeplearning.com Tasks in Computer Vision Lincoln 0.8 Washington 0.1 classification Jefferson 0.05 Obama 0.05 Input Image Pixel Representation - Regression: output variable takes continuous value - Classification: output variable takes class label. Can produce probability of belonging to a particular class 6.S191 Introduction to Deep Learning 1/29/19 introtodeeplearning.com High Level Feature Detection Let’s identify key features in each image category Nose, Wheels, Door, Eyes, License Plate, Windows, Mouth Headlights Steps 6.S191 Introduction to Deep Learning 1/29/19 introtodeeplearning.com Manual Feature Extraction Detect features Domain knowledge Define features to classify Problems? 6.S191 Introduction to Deep Learning 1/29/19 introtodeeplearning.com Manual Feature Extraction Detect features Domain knowledge Define features to classify 6.S191 Introduction to Deep Learning [2] 1/29/19 introtodeeplearning.com Manual Feature Extraction Detect features Domain knowledge Define features to classify 6.S191 Introduction
    [Show full text]
  • Controllable Person Image Synthesis with Attribute-Decomposed GAN
    Controllable Person Image Synthesis with Attribute-Decomposed GAN Yifang Men1, Yiming Mao2, Yuning Jiang2, Wei-Ying Ma2, Zhouhui Lian1∗ 1Wangxuan Institute of Computer Technology, Peking University, China 2Bytedance AI Lab Abstract Generated images with editable style codes This paper introduces the Attribute-Decomposed GAN, a novel generative model for controllable person image syn- thesis, which can produce realistic person images with de- sired human attributes (e.g., pose, head, upper clothes and pants) provided in various source inputs. The core idea of the proposed model is to embed human attributes into the Controllable Person Image Synthesis latent space as independent codes and thus achieve flexible Pose Code and continuous control of attributes via mixing and inter- Style Code polation operations in explicit style representations. Specif- Component Attributes ically, a new architecture consisting of two encoding path- Pose Attribute (keypoints) base head upper clothes pants ways with style block connections is proposed to decompose the original hard mapping into multiple more accessible subtasks. In source pathway, we further extract component layouts with an off-the-shelf human parser and feed them into a shared global texture encoder for decomposed latent codes. This strategy allows for the synthesis of more realis- Pose source Target pose Source 1 Source 2 Source 3 Source 4 tic output images and automatic separation of un-annotated Figure 1: Controllable person image synthesis with desired attributes. Experimental results demonstrate the proposed human attributes provided by multiple source images. Hu- method’s superiority over the state of the art in pose trans- man attributes including pose and component attributes are fer and its effectiveness in the brand-new task of component embedded into the latent space as the pose code and decom- attribute transfer.
    [Show full text]
  • Vision Based Self-Driving
    Direct Perception for Autonomous Driving Chenyi Chen DeepDriving • http://deepdriving.cs.princeton.edu/ Direct Perception estimate control Direct Perception (in Driving…) • Ordinary car detection: Find a car! It’s localized in the image by a red bounding box. • Direct perception: The car is in the right lane; 16 meters ahead • Which one is more helpful for driving task? Machine Learning What’s machine learning? • Example problem: face recognition Prof. K Prof. F Prof. P Prof. V Chenyi • Training data: a collection of images and labels (names) Who is this guy? • Evaluation criterion: correct labeling of new images What’s machine learning? • Example problem: scene classification road road sea mountain city • a few labeled training images What’s the label of this image? • goal to label yet unseen image Supervised Learning • System input: X • System model: fθ(), with parameter θ • System output: Ŷ= fθ(X), • Ground truth label: Y • Loss function: L(θ) • Learning rate: γ Why machine learning? • The world is very complicated • We don’t know the exact model/mechanism between input and output • Find an approximate (usually simplified) model between input and output through learning Deep Learning: A sub-field of machine learning Why deep learning? How do we detect a stop sign? It’s all about feature! Some Salient Features It’s a STOP sign!! Why deep learning? How does computer vision algorithm work? It’s all about feature! Pedestrian found!!! Why deep learning? • We believe driving is also related with certain features • Those features determine what action to take next • Salient features can be automatically detected and processed by deep learning algorithm Deep Convolutional Neural Network Deep Convolutional Neural Network (CNN) Convolutional layers Fully connected layers …… Deep feature, 4096D, float (4 bytes) Parallel computing Output, 14D, float (4 bytes) Input image, 280*210*3, int8 (1 byte) Figure courtesy and modified of Alex Krizhevsky, Ilya Sutskever, Geoffrey E.
    [Show full text]
  • Designing Neuromorphic Computing Systems with Memristor Devices
    Designing Neuromorphic Computing Systems with Memristor Devices by Amr Mahmoud Hassan Mahmoud B.Sc. in Electrical Engineering, Cairo University, 2011 M.Sc. in Engineering Physics, Cairo University, 2015 Submitted to the Graduate Faculty of the Swanson School of Engineering in partial fulfillment of the requirements for the degree of Doctor of Philosophy University of Pittsburgh 2019 UNIVERSITY OF PITTSBURGH SWANSON SCHOOL OF ENGINEERING This dissertation was presented by Amr Mahmoud Hassan Mahmoud It was defended on May 29th, 2019 and approved by Yiran Chen, Ph.D., Associate Professor Department of Electrical and Computer Engineering Amro El-Jaroudi, Ph.D., Associate Professor Department of Electrical and Computer Engineering Samuel J. Dickerson, Ph.D., Assistant Professor Department of Electrical and Computer Engineering Natasa Miskov-Zivanov, Ph.D., Assistant Professor Department of Electrical and Computer Engineering Bo Zeng, Ph.D., Assistant Professor Department of Industrial Engineering Dissertation Director: Yiran Chen, Ph.D., Associate Professor ii Copyright © by Amr Mahmoud Hassan Mahmoud 2019 iii Designing Neuromorphic Computing Systems with Memristor Devices Amr Mahmoud Hassan Mahmoud, PhD University of Pittsburgh, 2019 Deep Neural Networks (DNNs) have demonstrated fascinating performance in many real- world applications and achieved near human-level accuracy in computer vision, natural video prediction, and many different applications. However, DNNs consume a lot of processing power, especially if realized on General Purpose GPUs or CPUs, which make them unsuitable for low-power applications. On the other hand, neuromorphic computing systems are heavily investigated as a poten- tial substitute for traditional von Neumann systems in high-speed low-power applications. One way to implement neuromorphic systems is to use memristor crossbar arrays because of their small size, low power consumption, synaptic like behavior, and scalability.
    [Show full text]
  • CS4501: Introduction to Computer Vision Neural Networks (Nns) Artificial Neural Networks (Anns) Multi-Layer Perceptrons (Mlps Previous
    CS4501: Introduction to Computer Vision Neural Networks (NNs) Artificial Neural Networks (ANNs) Multi-layer Perceptrons (MLPs Previous • Softmax Classifier • Inference vs Training • Gradient Descent (GD) • Stochastic Gradient Descent (SGD) • mini-batch Stochastic Gradient Descent (SGD) • Max-Margin Classifier • Regression vs Classification • Issues with Generalization / Overfitting • Regularization / momentum Today’s Class Neural Networks • The Perceptron Model • The Multi-layer Perceptron (MLP) • Forward-pass in an MLP (Inference) • Backward-pass in an MLP (Backpropagation) Perceptron Model Frank Rosenblatt (1957) - Cornell University Activation function ": .: - .; "; 1, if ) .*"* + 0 > 0 ! " = $ *+, . ) 0, otherwise < " < .= "= More: https://en.wikipedia.org/wiki/Perceptron Perceptron Model Frank Rosenblatt (1957) - Cornell University ": !? .: - .; "; 1, if ) .*"* + 0 > 0 ! " = $ *+, . ) 0, otherwise < " < .= "= More: https://en.wikipedia.org/wiki/Perceptron Perceptron Model Frank Rosenblatt (1957) - Cornell University Activation function ": .: - .; "; 1, if ) .*"* + 0 > 0 ! " = $ *+, . ) 0, otherwise < " < .= "= More: https://en.wikipedia.org/wiki/Perceptron Activation Functions Step(x) Sigmoid(x) Tanh(x) ReLU(x) = max(0, x) Two-layer Multi-layer Perceptron (MLP) ”hidden" layer & Loss / Criterion ! '" " & ! '# # & )(" )" ' !$ $ & ! '% % & Linear Softmax $" = [$"& $"' $"( $")] !" = [1 0 0] !+" = [,- ,. ,/] 0- = 1-&$"& + 1-'$"' + 1-($"( + 1-)$") + 3- 0. = 1.&$"& + 1.'$"' + 1.($"( + 1.)$") + 3. 0/ = 1/&$"& + 1/'$"' + 1/($"( + 1/)$") +
    [Show full text]