Break Free, Taryn Southern’s AI-assisted Music Video

1 Break Free, Taryn Southern s AI-assisted Music VideoArtificial IntelligenceBy George Naugles for BUS550 with Minder Chen, Intelligence* machine learning *Deep learning *Neural Networks*Electronic (c)Copyright George Naugles, 2017, all rights IntelligenceWhat?So What? Now What? What? + Wires + Fiber + Code Mimic Brains ..with Neural Nets, big versions of thisTaking you where you ask ..still working on truck-evasion while !!!Fresh Direct feeds millions with AI, powered by gaming chips!Elemental AI Multi-core microprocessors like the 8 core NVDA chip in Playstation 4, ASICs=Application Specific Integrated Circuits, FPGA s = Field Programmable Gate Arrays.

2 All with divide and conquer coding/processing approaches ( )AI in John Gallaugher s Information Systems: A Manager s Guide to Harnessing Technology Remember the NVDA 3D graphics chips so important to VR and AR? Now there is a rise of similar but better chips specialized for machine learning and AI, enabling super-complex applications like blockchain tech and state of the art security. (IS:MGtHT ) FreshDirect model crushes costs that plague traditional Intelligence software, coupled with over 7 miles of fiber optic cables and sensors supports everything from baking the perfect baguette to verifying orders with accuracy.

3 ( ) (IS:MGtHT=our class text Information Systems: Manager s Guide to Harnessing Technology)AI Value PropositionAI = neural nets + deep machine learning , ..enable machine learning to do any task for which high-quality data can be gathered to enact a process of continuous improvement =>Business Process: collect and organize lots of data in formats your current and future AI computer can process, ..Which data? Any data you think may help your AI build a predictive model for your biz, subject to cost Networks, this instrument can learn Implication: Collect, Organize, & Mine AMHQDAP AMHQRDAP = As Much High Quality Relevant Data As Possible Today s AI software and hardware can write its own models by looking at reams of unstructured data to find predictive patterns and build predictive models from all the data we can feed them2018-2028 we ll manage neural nets,both human and machine .

4 Catering to YELP, FB, & TWTR Reviews in addition to price per share, which in turn is affected by data mining that taps real-time mobile internet quality data + AI guide our decision making as managers may determine which winner takes all in AI-powered 21st century business. Super-human customer-satisfaction capabilities capable of creating unprecedented customer loyalty and CLV characterize successes like AMZN, NFLX, GOOG, and NetBrainsFunction with neural networks classified as default, reward, affect, and are AI applications you can, as a manager, add to your team in order to gain powerful insights from data your human employees would find overwhelming.

5 Are some inputs AI currently helps us process? Vision data consisting of 25 to 30 frames per of data too great for any human to quantities of high quality data with predictive value that AI can discover and use to inform managers about the likely future with or without specific decisions .. to select and offer projections with unprecedented increasing number of Medical Diagnosis tasks and outcomes for judges where attorneys help clients decide Vision, which one is real?Argus 2 visual recognition enabled by AIIncreasing computational capability for computer,Microsoft s Visual Studio LifeCam with facial recognitionAI is what scientists are using to solve the most vexing problems, like climate models and projecting climate changeA robots teach themselves hand-eye coordination(start at 8:55) GoogleBrain is designing specialized computersAccelerates neural net computations at least one order of magnitude.

6 Does Linear Algebra extremely well. machine learning Is .. like making good wine Good data, like good grapes = essential Sebastian Thrun, Stanford Research Prof, founder of Udacity, team member for GoogleStreetView, GoogleGlass, Google GoogleX, Self-Driving Car Challenge, Mallone, Physicist, Udacity machine learning Specialist GOOD DATA MAKES GOOD MODELSLike good grapes make good wineIn Business, Intelligence Generalized learning , Reasoning, Problem Solving, & Language Intelligence involves: machine learning , computer vision, natural language processing, robotics, pattern recognition, & knowledge management for self-driving car control, recommendation engines, and optimization of just about any AI simulates the human brain, while weak AI immitates human behavior in ways that are not biomimicry.

7 Like DeepBlue a chess playing AI that considers thousands of moves before choosing one. Watson is more advanced, determines x% confidence level of answer. Google s AlphaGo was even more s better than Watson?Google s DeepLearning and GoogleBrain immitate brain structure, with artificial neurons, and neurological biomimicry to achieve greater processing capability, more creative capabilities, and more autonomy to learn how to perform as humans Networks = subsite of machine Decision-making based on data has depended on management information system and managers, AI has the potential to assist managers, and possibly replace them, making superhuman decisions.

8 There seems to be an AI arms race in a growing number of markets, allowing leaders to out-compete more primitive conventional firms with superhuman performance connection, creativity, and customer intimacy are seen as exclusively human realms for now, yet AI has the potential to outperform in those areas as Issues of Emerging AI Technologies*Modernizing to compete: AI Arms Race between Google, Intel, IBM, Amazon, Netflix and its competitors, and a wide range of start-ups writing code and creating optimized AI hardware.

9 *Google Cloud Platform and Amazon Web Services both offer Deep learning machine learning Artificial Intelligence mainframe capabilities that allow any start up to spin-up thousands of instances on seconds notice. This dramatically decreases barriers to entry for entrepreneurial startups using AI.*Will these AI applications escape the control of their designers, self-replicate on the internet, and create new software that has negative effects for one or more internet or mobile technology users in a future world full of IoT?

10 So What?How can we make money from superhuman artificial intelligence?Can you find Artificial Intelligence in this StarTrek image? Can humans hope to compete with AI?If we cannot beat AI s, how can we hire them and cooperate?350 AI researchers say MIT s McAfee is of their projections are actually coming true faster than expected: Language translation 2024,Essay writing 2026,Truck-driving 2027, Retail work 2031, Writing a best-seller 2049,Surgery 2053, and all human jobs wil be automated within 120 Neural Networks build layers of abstraction as part of learning processComplexity increases as the process is now open sourced.