Lecture 0: Introduction To Cognitive Computing And Deep Learning
UNIVERSITY OF JYVÄSKYLÄLecture 0: Introduction toCognitive Computing and Deep LearningTIES4911 Deep-Learning for Cognitive Computing for DevelopersSpring 2022by:Dr. Oleksiy KhriyenkoIT FacultyUniversity of Jyväskylä
UNIVERSITY OF JYVÄSKYLÄAcknowledgementI am grateful to all the creators/owners of the images that I found from Google and haveused in this presentation.21/01/2022TIES4911 – Lecture 02
UNIVERSITY OF JYVÄSKYLÄIBM Deep BlueIn 1996 and 1997 years, there was a pair of six-game chess matchesbetween world chess champion Garry Kasparov and an IBMsupercomputer called Deep Blue.https://en.wikipedia.org/wiki/Deep Blue versus Garry Kasparov Kasparov won the first match inPhiladelphia in 1996 (4:2). In 1997 in New York City, the firstcomputer program Deep Bluedefeated a world champion in amatchundertournamentregulations (3.5:2.5)21/01/2022TIES4911 – Lecture 03
UNIVERSITY OF JYVÄSKYLÄIBM WatsonWatson is an artificially intelligent cognitive computer system capableof processing large amounts of unstructured data and answering toqueries posed in natural language.https://www.youtube.com/watch?v Xcmh1LQB9IJeopardy! Challenge.In February 2011, IBM Watson made history competed against theworld’s best Jeopardy! Champions (Ken Jennings and Brad Rutter).https://www.youtube.com/watch?v P18EdAKuC1U21/01/2022TIES4911 – Lecture 04
UNIVERSITY OF JYVÄSKYLÄGoogle's AlphaGoGoogle's AlphaGo A.I. beats world's number one in html In October 2015, the distributedversion of AlphaGo defeated theEuropean Go champion Fan Hui(2-dan)(5:0) In March 2016, in a five-gamematch, the first time a computerGo program has beaten a 9-danprofessional Lee Sedol (one of thebest players)(4:1) In May 2017, in a three-gamematch, Global Champ Chineseplayer Ke Jie(9-dan), lost his firstgame against Google DeepMindcomputer program AlphaGo.AlphaGo ZeroIntroduces in October 2017, the latest evolution of AlphaGo is even more powerful and isarguably the strongest Go player in rning-scratchMinigo: an open-source implementation of the AlphaGoZero algorithm (https://github.com/tensorflow/minigo)In December 2017, DeepMind claimed that it generalized AlphaGo Zero's approach into a single AlphaZeroalgorithm. https://en.wikipedia.org/wiki/AlphaZeroMastering the game of Go without human 70.epdf21/01/2022TIES4911 – Lecture 05
UNIVERSITY OF JYVÄSKYLÄIntelligent RobotsAnArtificialSuperIntelligentRobotto study at collage, wants job, family, citizenship, etc.SophiawantsSophia is a social humanoid robot developed by Hong Kongbased company Hanson Robotics in 2015. In October 2017,the robot became a Saudi Arabian citizen, the first robot toreceive citizenship of any country. https://en.wikipedia.org/wiki/Sophia (robot) ow-about-sophia-worldsfirst-robot-citizen/ t-doing-now/ https://www.youtube.com/watch?v IUJaO6C-zTo https://www.youtube.com/watch?v 8MjIU4eq A https://www.youtube.com/watch?v fLvL7uqrMVc https://www.youtube.com/watch?v S5t6K9iwcdw https://www.youtube.com/watch?v XwRXv80AUTs 21/01/2022TIES4911 – Lecture 06
UNIVERSITY OF JYVÄSKYLÄIntelligent RobotsMicrosoft and Alibaba AI programs beat humans in Stanford readingcomprehension test for 1st timeMachines can already outplay us in chess, poker and othergames, and now they are becoming better readers as well.AI programs from both Microsoft and Alibaba outperformedhumans in the beginning of January 2018 on a Crowdworkers” scraped more than 500 Wikipedia articles toproduce more than 100,000 question-and-answer sets for thetest.Here’s a sample question: “What year did Genghis Khan die?”(Spoiler alert: It’s 1227.)“This is the first time that a machine has outperformed humanson such a test,” Alibaba said in a statement.Microsoft’s score of 82.6 and Alibaba’s grade of 82.4 beat outthe human standard of 82.3. Other notable AI programsparticipating in the test and closing in on beating human scorescome from the Allen Institute for Artificial Intelligence, Tencent,Salesforce and others. e/21/01/2022TIES4911 – Lecture 07
UNIVERSITY OF JYVÄSKYLÄIntelligent RobotsIntelligent Virtual Assistants and Chat Bots Amazon AlexaGoogle HomeApple Sirihttps://en.wikipedia.org/wiki/Amazon e.com/siri/IBM Watson AssistantMicrosoft CortanaFacebook elopers.facebook.com/21/01/2022TIES4911 – Lecture 08
UNIVERSITY OF JYVÄSKYLÄSelf-Driving CarsNVIDIA and Bosch Announce AISelf-Driving Car osch/21/01/2022TIES4911 – Lecture 09
UNIVERSITY OF JYVÄSKYLÄIntelligent RobotsBOSTON DYNAMICS is a world leader in mobile robots, tackling some of the toughestrobotics challenges. https://www.bostondynamics.comIt combines the principles of dynamic control and balance withsophisticated mechanical designs, cutting-edge electronics, and nextgeneration software for high-performance robots equipped withperception, navigation, and intelligence. Boston Dynamics has anextraordinary and fast-growing technical team of engineers andscientists who seamlessly combine advanced analytical thinking withbold engineering and boots-in-the-mud practicality. https://www.youtube.com/user/BostonDynamics https://www.youtube.com/watch?v NR32ULxbjYc21/01/2022TIES4911 – Lecture 010
UNIVERSITY OF JYVÄSKYLÄAutonomous DeliveryRobot delivery dogs deployed by self-driving cars �s hope you’re notafraid of dogs, because ifContinental gets its way,autonomous robot dogsare going to be deliveringyour packages.21/01/2022TIES4911 – Lecture 011
UNIVERSITY OF JYVÄSKYLÄAutonomous DeliveryAgility’s two-legged robot Digit is for sale and Ford isthe first customer customer/21/01/2022TIES4911 – Lecture 012
UNIVERSITY OF JYVÄSKYLÄLogistics RobotsIndoor Logistics Robots Solteq Indoor Logistics Robots moves autonomouslycarrying loads up to 100/550 kilograms. The robotcan be integrated with other systems, such aselevators, used in the premises. 21/01/2022TIES4911 – Lecture 013
UNIVERSITY OF JYVÄSKYLÄIntelligent RobotsInventory Robots in Retailing obot https://www.youtube.com/watch?v nda7bTLNcQs21/01/2022TIES4911 – Lecture 014
UNIVERSITY OF JYVÄSKYLÄIntelligent RobotsRobots in Education 21/01/2022TIES4911 – Lecture 015
UNIVERSITY OF JYVÄSKYLÄIntelligent RobotsRobots in Service 21/01/2022TIES4911 – Lecture 016
UNIVERSITY OF JYVÄSKYLÄNeuromorphic ComputingNeuromorphic Computing: The Future of AI and ComputingXeon Scalable Processors and Intel Nervana Neural Network Processors (NNP) Redefine AI SiliconFamilyofIntel lligence.html ocessors-for-deep-learning-training processors-nnp-redefine-aisilicon/?utm source ISTV&utm medium Video&utm campaign ISTV2018 ISTV1802 01&utm content AI News https://www.youtube.com/watch?v zEzm-rMwyVo https://www.youtube.com/watch?v ej9-sGj iHwIntel Loihi AI chipA new AI chip under development at Intelis taking its inspiration from the human brain in an attempt toovercome technological hurdles. This New Self-Learning ChipPromises to Accelerate Artificial Intelligence https://www.youtube.com/watch?v EgCRwZw4p8chttps://www.youtube.com/watch?v tificial-intelligence/21/01/2022TIES4911 – Lecture 017
UNIVERSITY OF JYVÄSKYLÄ?!“Slaughterbots”is seven minutes and forty-seven seconds of sheer horrordesigned to be a fictional warning against a future full of killer robots.This fictional video about AI-powered weapons makes The Terminator look like a Disney film tm content buffer215b3&utm medium social&utm source facebook.com&utm campaign buffer https://www.youtube.com/watch?v 9CO6M2HsoIA http://autonomousweapons.org21/01/2022TIES4911 – Lecture 018
UNIVERSITY OF JYVÄSKYLÄ?!Something went wrong https://www.youtube.com/watch?v y3RIHnK0 NEhttps://www.youtube.com/watch?v ZoemTySxFso21/01/2022TIES4911 – Lecture 019
UNIVERSITY OF JYVÄSKYLÄ?!Deep Fake https://www.youtube.com/watch?v 78L6I6vsfrUhttps://www.youtube.com/watch?v Wm3squcz7Aw21/01/2022TIES4911 – Lecture 020
UNIVERSITY OF JYVÄSKYLÄDeep Fake Teenager's AI Project for Detecting DeepfakeVideos Wins AwardHis software has over 150,000lines of code and is ten timesfaster than the current wardBT Young Scientist & Technologist of the year 2021 - Greg Tarrhttps://www.youtube.com/watch?v pFYE6O4rw24https://www.youtube.com/watch?v rHQtst-WlbkTarr was able to make significant improvements in speed and efficiency when compared tothe current state-of-the-art best model without sacrificing its accuracy.Microsoft novel deep fake detection ation21/01/2022TIES4911 – Lecture 021
UNIVERSITY OF JYVÄSKYLÄCognitive ComputingCognitive Computing is a new type ofcomputing with the goal of more accuratemodels of how the human brain/mindsenses, reasons, and responds to stimulus.https://en.wikipedia.org/wiki/Cognitive computingCognitive Computing based systems are “systems that learn at scale,reason with purpose and interact with humans naturally.” (IBM)Types of cognitive technologies: Machine learningNatural language processingSpeech recognitionComputer visionInsights generation from dataSentiment analysisEtc.21/01/2022TIES4911 – Lecture 022
UNIVERSITY OF JYVÄSKYLÄCognitive ComputingOn the way towards Cognitive Computing, smartsystems adopt key elements of cognitive computing Expanding the boundaries of human cognition extends a capability of a human to reason, think deeply,recognize objects and sounds, manipulate and manage hugeamount of data (not only to search in big volume, but makedecisions on top of it). More natural human-computer interaction applies more natural interaction and engagement withcomputers via more general speech and natural languagecommunication with the system, as well as, use of infographicsand visual data representation techniques. Use of Learning helps to design personalized and adaptable systems able toconstantly learn and evolve based on feedback from usedinteraction applying machine learning, statistics, etc.21/01/2022TIES4911 – Lecture 023
UNIVERSITY OF JYVÄSKYLÄDeep LearningMachine Learning- process of training a machine to create a model and use it fordecision making.Deep Learning- is part of a broader familyof machine learning methods based on learningdata representations, as opposed to taskspecific algorithms.(Wikipedia)“The analogy to deep learning is that the rocket engine is the deeplearning models and the fuel is the huge amounts of data we canfeed to these algorithms.” (Andrew Ng)Deep Learning vs. Machine Learning – the essential differences you need to rning-tutorialhttps://www.youtube.com/watch?v vehXkgG3YcU21/01/2022TIES4911 – Lecture 024
UNIVERSITY OF JYVÄSKYLÄDeep LearningDL use-cases Computer/Machine Vision image classification and automatic taggingobject recognition in the imagevideo recognition Speech recognitionand generation Text processing fact extractionmachine translationsentiment analysischaracter level text processingdocument classification Decision making Etc.21/01/2022TIES4911 – Lecture 025
UNIVERSITY OF JYVÄSKYLÄDeep LearningDL application domains Medical cancer detectiondrug discoveryradiology (CNN based detection of tumor and cancerfrom MRI, fMRI, EKG, and CT scans) Finance market, trading and investment predictionscustomer segmentation in advertisingfraud detection Agriculture problematic environmental conditions detection basedon satellite feeds and sensor dataSmart CitiesTraffic &TransportationGamingMusic and ArtRoboticsEducationService domainsEtc.21/01/2022TIES4911 – Lecture 026
UNIVERSITY OF JYVÄSKYLÄDeep LearningDL application domains Forecasting Waves with Deep g-Waves-with-Deep-Learning.aspx CheXNet: Radiologist-Level PneumoniaDetection on Chest X-Rays with DeepLearning / ts.html Face It cially-intelligenthairstylist?utm source taboola&utm medium referral&utm campaign AI Student ASMO Q4 2017 Media Buy21/01/2022TIES4911 – Lecture 027
UNIVERSITY OF JYVÄSKYLÄDeep LearningDL most known researchers Andrew NgGeoff HintonYann LeCunYoshua BengioAndrej Karpathy Comprehensive Neural Networklearning materials Michael Nielsen's book:http://neuralnetworksanddeeplearning.com/ Andrew Ng's 2TIES4911 – Lecture 028
UNIVERSITY OF JYVÄSKYLÄDeep LearningDL companies (big players)BOSCHsupports development of autonomous vehiclesAppleGoogleis actively investing into self-driving carsbought DeepMind for 400 Millionand provides Cognitive Computing servicesNVIDIAimproved DL hardware with GPUsIntelintroduced Intel Nervana NN processor andnew Loihi Self-Learning Chip, and improvesDL toolkits, frameworks and algorithms IBM WatsonCognitive Computing servicesFacebook AIDL frameworks and toolsMicrosoft AzureAmazonCognitive services and toolkitToyotaAWS Cloud Cognitive Serviceshas invested 1 Billion into AI research21/01/2022TIES4911 – Lecture 029
UNIVERSITY OF JYVÄSKYLÄDeep Learning FrameworksIntelRelevant -frameworks fig1 34910871421/01/2022TIES4911 – Lecture 0delivers fast and scalable AI in production,streamlines AI inference by enabling teamsdeploy trained AI models from any framework(TensorFlow, NVIDIA TensorRT , PyTorch,ONNX, XGBoost, Python, custom and more onany GPU- or CPU-based infrastructure (cloud,data center, or ference-server30
UNIVERSITY OF JYVÄSKYLÄ TensorFlowDeep Learning Frameworksby Google seems to be the most used deep learning framework so far. In 2017, Google hasintroduced TensorFlow Lite - is a lightweight solution for mobile and embedded devices, enabling on-device machinelearning inference and supporting hardware acceleration with the Android Neural Networks API. Also,TensorFlowintroduced TensorFlow.js that allows development and training models using Javascript. https://www.tensorflow.org/ PyTorchwas introduced by Facebook in January 2017 and already started to gain popularity. The main driversbehind the popularity are GPU acceleration, the efficient usages of memory and the use of dynamic computationalgraphs (“define by run” instead of the traditional “define and run”). Similarly to TF, PyTorch Mobile was introduced tohelp users launch models on mobile devices. http://pytorch.org/ Caffe2framework has been launched by Facebook in 2017 as the successor of the well known and still extremelypopular Caffe framework. https://caffe2.ai/ and http://caffe.berkeleyvision.org/ MXNet is Apache library for deep learning supported by Microsoft and Amazon. It supports many languages, fromC to Python, JavaScript, Go, and, indeed, R. https://mxnet.apache.org/ CNTKdeep learning framework developed by Microsoft. The framework was renamed to the Microsoft CognitiveToolkit with officially launched the 2.0 version in 2017. https://www.microsoft.com/en-us/cognitive-toolkit/ Torchis a scientific computing framework for LuaJIT with wide support for machine learning algorithms that putsGPUs first. http://torch.ch/ Theanois a Python library that allows to efficiently define, optimize, and evaluate mathematical expressionsinvolving multi-dimensional arrays. http://deeplearning.net/software/theano/ Deeplearning4j is an open-Source, distributed, deep Learning Library for the JVM. https://deeplearning4j.org/ Chainer is a Python-based deep learning framework aiming at flexibility. IntelChainer is Optimized-Chainer for Intel Architectures. https://chainer.org/, https://github.com/chainer/chainer and https://github.com/intel/chainerNeon is a deep learning (Python-based and optimized for Intel architecture) framework designed for ease of use andextensibility on modern deep neural networks, such as AlexNet, Visual Geometry Group (VGG), and emy/frameworks/neon/ and https://ai.intel.com/neon/Intel provides Framework Optimizations for faster training of deepneural networks on Intel architecture /2022TIES4911 – Lecture 031
UNIVERSITY OF JYVÄSKYLÄDeep Learning FrameworksInterfaces that are wrapped around one or multiple frameworks: Kerasis the most well know and widely used interface for deep learning. This high-level Python based deeplearning API is created by a deep learning researcher at Google - François Chollet. Google announced in 2017 thatKeras has been chosen to serve as the high-level API of TensorFlow and will be included in the next TensorFlowrelease. Next to TensorFlow, Keras can also use Theano or CNTK as backend. https://keras.io/ PyTorch Lightningis the Keras of PyTorch that has been released to ease and shorten the process ofimplementing neural networks easier Gluon is an open source high-level Python deep learning interface which allows developers to more easily andquickly build machine learning models, without compromising performance was jointly announced by Microsoft andAmazon’s AWS in October 2017. Interface wraps MXNet and soon it will also include Microsoft’s x.html Eager executionfor TensorFlow, introduced in October 2017, is an imperative “define-by-run” interfacewhere operations are executed immediately as they are called from Python. With this launch, Google hoped to winback the users that fell in love with PyTorch and it’s dynamic graph. For TensorFlow 2.x it is default functionality rib/eager/python/g3doc/guide.md Sonnetis a library (by DeepMind) built on top of TensorFlow for building complex neural onnet/ H2O Deep Wateris a H2O for GPU Enabled Deep Learning on all data types integrating with TensorFlow,MXNet and Caffe. Deep Water brings all these frameworks together under the same user interfaces as the H2Oplatform. Now, in addition to the original H2O Deep Learning algorithm, users can access TensorFlow, MXNet andCaffe backends in H2O, and build complex deep networks. https://www.h2o.ai/deep-water/ONNX (Open Neural Network Exchange) (is launched by Microsoft, AWS, and Facebook amongst others)is anopen format to represent deep learning models that allows users to more easily move models between different frameworks (V1 isreleased in December 2017). ONNX enables models to be trained in one framework and transferred to another for inference. ONNXmodels are currently supported in Caffe2, Microsoft Cognitive Toolkit, MXNet, and PyTorch, and there are connectors for many othercommon frameworks and libraries (community already added a converter for TensorFlow as well). https://onnx.ai/ andhttps://github.com/onnx/onnx21/01/2022TIES4911 – Lecture 032
UNIVERSITY OF JYVÄSKYLÄDeep Learning Platforms H2O.ai is a machine learning platform. https://www.h2o.ai/ Spark is a fast and general engine for large-scale data processing. http://spark.apache.org/ PlaidML - open source portable deep learning engine from https://github.com/plaidml/plaidml A.I. Model- a common machine learning tool for all frameworks by atformthree major tools for developers: Azure Machine Learning Experimentation ServiceAzure Machine Learning WorkbenchAzure Machine Learning Model Management Service. IBM Watson and corresponding Watson Data Platform and IBM Services Platform.https://www.ibm.com/watson/ Google Cloud AImodelsandaprovides modern machine learning services, with / Intel AI DevCloud for Intel AI Academy s/devcloud21/01/2022TIES4911 – Lecture 033
UNIVERSITY OF JYVÄSKYLÄCognitive Computing ServicesGoogle cloud.google.com/speech/ IBM tural-language-understanding/ Microsoft us/services/cognitive-services/text-analytics/ levant 1 – Lecture 034
Cognitive Computing based systems are "systems that learn at scale, reason with purpose and interact with humans naturally." (IBM) Types of cognitive technologies: Machine learning Natural language processing Speech recognition Computer vision Insights generation from data Sentiment analysis Etc. 21 Cognitive Computing
Introduction of Chemical Reaction Engineering Introduction about Chemical Engineering 0:31:15 0:31:09. Lecture 14 Lecture 15 Lecture 16 Lecture 17 Lecture 18 Lecture 19 Lecture 20 Lecture 21 Lecture 22 Lecture 23 Lecture 24 Lecture 25 Lecture 26 Lecture 27 Lecture 28 Lecture
Lecture 1: A Beginner's Guide Lecture 2: Introduction to Programming Lecture 3: Introduction to C, structure of C programming Lecture 4: Elements of C Lecture 5: Variables, Statements, Expressions Lecture 6: Input-Output in C Lecture 7: Formatted Input-Output Lecture 8: Operators Lecture 9: Operators continued
Lecture 1: Introduction and Orientation. Lecture 2: Overview of Electronic Materials . Lecture 3: Free electron Fermi gas . Lecture 4: Energy bands . Lecture 5: Carrier Concentration in Semiconductors . Lecture 6: Shallow dopants and Deep -level traps . Lecture 7: Silicon Materials . Lecture 8: Oxidation. Lecture
TOEFL Listening Lecture 35 184 TOEFL Listening Lecture 36 189 TOEFL Listening Lecture 37 194 TOEFL Listening Lecture 38 199 TOEFL Listening Lecture 39 204 TOEFL Listening Lecture 40 209 TOEFL Listening Lecture 41 214 TOEFL Listening Lecture 42 219 TOEFL Listening Lecture 43 225 COPYRIGHT 2016
Partial Di erential Equations MSO-203-B T. Muthukumar tmk@iitk.ac.in November 14, 2019 T. Muthukumar tmk@iitk.ac.in Partial Di erential EquationsMSO-203-B November 14, 2019 1/193 1 First Week Lecture One Lecture Two Lecture Three Lecture Four 2 Second Week Lecture Five Lecture Six 3 Third Week Lecture Seven Lecture Eight 4 Fourth Week Lecture .
Introduction to Quantum Field Theory for Mathematicians Lecture notes for Math 273, Stanford, Fall 2018 Sourav Chatterjee (Based on a forthcoming textbook by Michel Talagrand) Contents Lecture 1. Introduction 1 Lecture 2. The postulates of quantum mechanics 5 Lecture 3. Position and momentum operators 9 Lecture 4. Time evolution 13 Lecture 5. Many particle states 19 Lecture 6. Bosonic Fock .
The Allen Cognitive Level Screen (ACLS) can help you identify the Allen Cognitive Levels of clients with Alzheimer’s disease, dementia, and other cognitive disabilities. Also referred to as the leather lacing tool, this cognitive assessment tool measures global cognitive processing capaci
MEDICAL RENAL PHYSIOLOGY (2 credit hours) Lecture 1: Introduction to Renal Physiology Lecture 2: General Functions of the Kidney, Renal Anatomy Lecture 3: Clearance I Lecture 4: Clearance II Problem Set 1: Clearance Lecture 5: Renal Hemodynamics I Lecture 6: Renal Hemodynamics II Lecture 7: Renal Hemodynam
