MIST5720 Final
Driving Forces Behind AI
-Processing vast amounts of data using powerful new machine learning strategies requires an equally powerful infrastructure and platform -High Bandwidth Services Allow for the speedy collection/transfer of data -Scalable, reliable, and powerful computational and storage abilities Cloud Computing provides a platform that is optimal for developing easily deployable and integratable AI solutions
Cons
Cybersecurity risks Data overload Assessing quality of data Need for culture change Rapid technology change Compliance
Data Security
Data is crucial to the operations of a company. Who? Where? How? Hadoop and NoSQL Concerns: Encryption failure User access control Hackers Vulnerability to fake data generation
Traditional Data vs. Big Data
-Traditional data uses fixed format or fields in a file, which means it is based on structured data -Big data uses the semi-structured and unstructured data and improves the variety of the data gathered from different sources like customers, audience or subscribers. After the collection, big data transforms it into knowledge based information
Self Driving Cars
-Uber announced they are purchasing 24,000 self driving Volvos -Tesla has designed an advance self driving tractor trailer and UPS has ordered 125 -Tesla is about to release one of the fastest cars on the road, and yes it will have a self driving option
What is Cloud Computing?
-Using a network of servers hosted on the Internet to store, manage, and process data "Off campus" IT infrastructure -Cloud capabilities used to increase efficiency in organizations *Cuts in operational costs *Automation *Connectivity *Accessibility -Companies offering these computing services are called cloud providers Charge for cloud computing services based on usage
Supervised Machine Learning
A machine learning algorithm is trained to correctly respond to questions related to feature vectors. To train an algorithm, the machine is fed a set of feature vectors and an associated label. Labels are typically provided by a human annotator, and represent the right "answer" to a given question. The learning algorithm analyzes feature vectors and their correct labels to find internal structures and relationships between them
The Backbone of Modern Innovation
AI enables high-powered innovation across multiple fields of technology -Provides enhanced speed and accuracy in algorithm creation and tuning -Provides an affordable and more effective way to process data -Makes the user more competitive in the market It is already heavily relied upon in many areas -IoT -Mechatronics -3D Visualization
Applications:
Entertainment Military Healthcare
What is Machine Learning?
Method of data analysis that automates analytical model building You see, you can get AI without using machine learning, but this would require building millions of lines of codes with complex rules and decision-trees
Predictive Analytics Using Big Data
Predictive analytics is one technique that is used commonly to analyze big data sets
Benefits of IoT
Reduce Maintenance Costs Safety and Protection Improves Convenience Access to Information New Revenue Better Customer Service and Support Business Opportunities Efficiency
Ways to collect big data
Sensors Devices Video/Audio Applications Web Social Media Transactional Loyalty/Rewards Cards Email
The 3 V's of Big Data
Volume Variety Velocity
Mixed Reality (MR)
A blend of the physical and digital worlds in which users may interact with digital and real-world objects while maintaining presence in the real world
PROS/CONS
Health Concerns: Most VR systems come with consumer warnings including: seizures; developmental issues in children; trip-and-fall collision warnings; discomfort; repetitive stress injury; and interference with medical devices. Some users may even experience twitches, seizures or blackouts while using VR headsets even if they don't have a history of epilepsy
Scale and Reimage Job Training and Workforce Skills Development
Prepare now for what is later to come Address potential skill gaps Investing now for a business priority that could become urgent later
ARCHITECTURE
Printing houses out of concrete New Story + ICON project in El Salvador A home like this can be built in less than 24 hours with 3D printing For under $4,000 New story - a housing charity organization and ICON - a construction tech company have partnered together - END GLOBAL HOMELESSNESS
IS THERE ANY BIAS?
ProPublica in 2016 wrote an article claiming bias in the algorithms Claimed the COMPAS program to be racial bias says that African Americans were being wrongly rated for likelihood of committing a new crime once incarcerated Many scholars looked at the data and said that because of the nature of the data even though the algorithm might not account for race it still can be bias. There is not conclusive proof either way because the AI is still learning to better predict future criminals
Alan Turing
a young British polymath who explored the mathematical possibility of artificial intelligence. Turing suggested that humans use available information as well as reason in order to solve problems and make decisions, so why can't machines do the same thing?
Potential Risks and Challenges of Cloud:What are Pro's and Con's?
Advantages Faster start-up and implementation of von services Needs-based scaling (On-Demand) Transparent cost control Low risk of investment Mobile access possible (Consumerization) Collaboration is enabled globally Focus on the core business competence Reduced operative risk Easy and conveniently accessible Disadvantages Dependence on provider Low adjustments possibilities Low security standards SLA partially non-transparent Standardization instead of innovation Requires adjustments for system change Data security & encryption Use requires coordination
Opportunities: Voting
Australia Post is currently looking into blockchain voting ● Allows a location agnostic, "tamper proof" system ● May not prevent all of voting fraud
Time Variance
Big data is constantly changing, while traditional data is where historical data is almost equally valid as data today. Example: your most recent website view will always be changing considering how often we are browsing the web. But basic information on historical figures such as president's names and birthdates are always going to be valid today.
WHAT ABOUT FOOD?
Columbia Professor Hod Lipson - The future of this technology will be food. "I think food printing is to 3D Printing what gaming was for computers."
IaaS: Infrastructure as a Service
Companies rent computing resources from provider on a pay-per-use basis Servers, storage, networks, infrastructure, OS IaaS provider handles all maintenance tasks, updates and backups Customers don't have to buy hardware / employ expert management Considered as the fundamental building block, offering services via APIs Most flexible cloud computing model, automated, highly scalable deployment with full control over the infrastructure → Main area of application: development of PaaS, SaaS, Web-scale apps
Technical Issues and Solutions
Costs and Constraints Devices Behaving Badly *Networking Resources Legacy Issues Configuration *Large Scale rollout Software Updates *Inherent tech issues Adequate Research and Development *Advance technology Harmonize Standards *Hold producers to a standard
SaaS: Software as a Service
Describes a way of delivering applications via the Internet Accessible right from a web browser, no special HW / SW required SaaS provider takes care of maintenance and updates Software management and deployment is moved to 3rd party services Use of SaaS allows users to reduce software ownership costs → no technical staff required to install, maintain and upgrade Software Provided on a subscription model, pay per use with high scalability
Connectivity Model
Device-to-Device Device-to-Cloud Device-to Gateway Back-End Data-Sharing
Start Taxing Businesses
Discourage Businesses Positive impact the decline rate of human jobs Weigh the costs even more Unpredictable because of growth
Creating and Deploying Machine Learning
Do your research and create a plan Choose your programming language Collect data, clean it, categorize it Choose the appropriate Machine Learning Methodology Choose the best algorithm Tune the algorithm Present Results (Somehow)
Find a Balance
Don't rely solely on AI, but integrate a balance that allows it to enhance the ability to serve clients
Hype Cycle
Gartner dropped Big Data from the Hype Cycle in 2015 It was near the "Peak of Inflated Expectations" in 2013 and slipped down into the "Trough of Disillusionment" only a year later Why was it dropped? It has become prevalent across many hype cycles It touches so many aspects of our interactions with computers
Google lawsuit
Google has been accused of illegally tracking millions of iphone and android users when they turn off "Location history". Google is actively monitoring and storing the location of people who choose not to be tracked. This violates their privacy terms.
REGULATIONS
Government regulated Add these now to slow down growth and develop general governing rules Reason they are not set now: there is no consensus on what it is/what it can do Complex Constantly changing
NEW FUTURE OUTLOOK
Healthcare providers, public sector, banking and securities, communications, media and services, retail, and wholesale trade will benefit from AI without ever suffering annual net job loss. Manufacturing and transportation are disproportionately big contributors to job losses: 938,000 manufacturing jobs will be eliminated due to AI by year-end 2019. Global IT services firms will have massive job churn in 2018; adding 100,000 jobs and dropping 80,000.
Privacy
How much do you know about yourself? Do companies that use big data know more than you? Companies like Facebook and Google know everything about you Target and smart advertising. Big data administrators may decide to mine data without permission or notification. Very few laws preventing or regulating data mining.
What is IoT
Internet of Things (IoT) is the collection of devices that can communicate over the Internet with each other or with a control console in order to affect and monitor the real world. IoT devices might be labeled as smart devices or smart home equipment.
APPLICATIONS
Medical Industrial Advertising Commercial Education Tourism Entertainment
CONS
Most AR, including those for advertising, requires a custom application to be downloaded to use Technology is still maturing, computer systems have only just become powerful enough for widespread adoption Invades privacy
Augmented Reality (AR)
Overlaying digital information onto the physical world with no or very limited interaction with virtual objects
WHAT ARE THE ADVANTAGES AND DISADVANTAGES?
PROS: Easy prototyping Fast production Decreased costs Design complexity CONS: Manufacturing jobs decrease Dangerous objects created with little regulation Not as suitable for mass manufacturing
IoT Economic Impacts
Payments Insurance Lean operations / Bottom lines Real time decisions Automation Tech industry growth 2014 $900 billion Projected 2024 $4.3 trillion
Veracity
Quality of data. Small data contains less noise as data collected in a controlled manner. While the quality of big data is not guaranteed. Rigorous data validation is required before processing big data.
Introduction to Cloud Computing and SaaS
SaaS as a part of Cloud Computing → Application provided via, external IT service provider provides software infrastructure and platforms, receiving a fee for the service Payment Payment depends on the usage (per month/ volume/ functionality / transaction Different contract models, flexibly adaptable to user's needs Implications for Entire Industries Revolutionary business models enabled through SaaS and Cloud support Savings for enterprises to concentrate on core competencies Process optimization & efficiency improvements through adequate capacities
Data source
Small data typically comes from traditional systems like CRM's ledgers, and payment transactions. But big data can be purchased and collected from new sources such as clickstreams from websites, GPS systems, social media, and more.
What is the Cloud?
Software and services that runs on the Internet Not on user's personal device Store data by accessing cloud-based servers using the internet "Unlimited" storage space Accessible on all devices 24/7 Where the hardware and software is located and how it all works is not important to user Somewhere up in the "clouds" Most cloud services can be accessed through web browsers Some companies offer dedicated applications
BRIEF HISTORY
Traditional manufacturing: subtractive processes -- considerable limitations Additive manufacturing much more efficient usage started out with plastic systems in the mid 80s but now has taken off with metals We can now easily 3D print with metal
expert systems
made by Edward Feigenbaum which mimicked the decision making process of a human expert The program would ask an expert in a field how to respond in a given situation, and once this was learned for virtually every situation, non-experts could receive advice from that program.
3D PRINTING
manufacturing process in which material is laid down, layer by layer, to form a three-dimensional object virtual → physical object AKA Rapid Prototyping, stereolithography, architectural modeling or additive manufacturing Requires digital file - STL files most common Design, slice, print
Opportunities: Healthcare
● Small early minority are piloting blockchain technology ● Data sharing between health organizations currently fragmented and inefficient ● Fast, secure, and decentralized data exchange
Opportunities: Property Rights
● Smart contracts ● Ethereum platform was created to enable and enforce smart contracts ● Automate the execution of contracts ● Speed is a current limitation
Impacts from Self Driving Cars
-People will not own cars, instead just order rides -Parking decks will become no more, instead they will be converted -Traffic will decrease because there will be much more carpooling -Battery replacement stations will replace gas stations -Development will change, there will be no need for driveways and parking spots
AI in Criminal Justice System
-Today, AI is used to set bail amount for suspected criminals and the punishment for convicted criminals -North Pointe's COMPAS is an example of this in which it helps to speed up judge's decisions, predict likelihood of new crimes that will be committed, and ultimately lead Judges to more appropriate decisions. -The algorithm is objective, race, gender, and geography are not used in its calculations and it learns from new data everyday. -This process can ultimately save people who are wrongly accused get out of jail quicker, and those who will most likely to not commit a crime be let out of prison quicker, thus reducing taxpayer waste.
HOW OLD IS 3D PRINTING?
1980 -Rapid Prototyping (RP) technologies 1984: Charles Hull stereolithography
What is Blockchain?
3 "pieces" to blockchain 1. Public, electronic ledger 2. Unchangeable/immutable 3. Consensus-based permission ● Made up of "blocks" and hashes ● No central authority ● No deleting entries
Computing Machinery and Intelligence
1950 paper where he discussed how to build intelligent machines and how to test their intelligence.
FLAWS
Changing Requirements Not adaptable on the spot Destructive
Hype Cycle
GARTNER HYPE CYCLE The 3D printing (3DP) retail marketing is becoming more vibrant The emegrence of 4d printing adds dimension of transformation - like self folding origami
POSITIVE GROWTH AHEAD
Gartner is confident about the positive effect of AI on jobs. The main contributor to the net job growth is AI augmentation — a combination of human and artificial intelligence, where both complement each other.
The Technology Behind IoT
RFID (radio-frequency identification) NFC (near-field communication) Low-Energy Bluetooth Low-Energy Wireless Radio Protocols WiFi-Direct LTE-Advanced
Unsupervised Machine Learning
The algorithm is programmed to predict answers without human labeling, or even questions. Rather than predetermine labels or what the results should be, unsupervised learning harnesses massive data sets and processing power to discover previously unknown correlations.
Data Location
Traditional data typically uses very centralized and local servers, but big data is stored in external file systems and in the cloud.
HOUSEHOLDS
Will create a weapon at every computer -- anyone can download digital blueprint Print a gun in your living room? What could possibly go wrong? Can be made without serial numbers - no background checks, completely untraceable We live in an era where you can pirate a Drake Album so how can you stop the flow of data?
Three Biggest Common Misconceptions
Artificial Intelligence: Will take over all of our jobs Cannot outsmart humans because they created it Just suddenly happened
Application of blockchain in finance
Bank payment transactions • Currency • Etherium • Ripple • BitCoin
WHAT IS AI GOOD FOR?
Better Automated Tasks and Processes Enhanced Data Collection and Analytical Power Easier Access to Previously Unknown/ Undiscovered Territory Improved Reliability
Pros
Better decision-making Increased productivity Reduced costs Improved customer services Increased revenue Faster speed to market Greater innovation
Value
Big data goes beyond analysis and reporting. It allows for complex data mining for prediction, recommendation, and pattern finding which is much more valuable for a company. If you're going to have so much data, you need to know how it is useful for you and what you can do with it.
IoT for Business
Business/Manufacturing Healthcare Retail Security Transportation
Security Concerns
Cloud Attacks . Device Attacks:
Hype Cycle of Cloud Computing?
Cloud computing has now reached the Slope of Enlightenment Shift from hype to inevitable mainstream acceptance Shift from unrealistic promises that cloud will transform everything Shift from pessimistic view that it didn't deliver Now a pragmatic approach that views cloud as a definite part of the approaching future
Predictive Analytics can be used in various arenas:
Identifying fraud Optimizing marketing Improving operations
Big Data Future Trend
In-Memory Technology Faster Processing SAP, IBM, PIvotal In-Memory database vendors Machine Learning Allowing computers to learn new things on their own Big Data Intelligent Apps Personalization and better services. Suggestion Engine like google, netflix, youtube Intelligent Security Cyberattacks attempts
What is Artificial Intelligence?
Machines that can perform tasks that are characteristic of human intelligence planning, understanding language, recognizing objects and sounds, learning, and problem solving.
THE FUTURE
VR's Main Applications: Will VR be used primarily in entertainment or in the health industry. Ethical Dilemmas: Is pushing the boundaries of the human mind dangerous to the development of the brain and how should video game developers adapt their content for younger and more vulnerable audiences? Should video game and content developers be required to take more effective steps than just "warning" of potentially harmful content just like the ESRB attempts to keep younger audiences from purchasing violent video games. The VR Industry's move towards the consumer market: The price of a highly capable VR system has decreased dramatically from thousands of dollars a few years ago to a few hundred dollars today. Some of the most popular headsets include the Sony Playstation VR ($250), the HTC Vive ($500), the Oculus Go ($200), the Samsung Gear VR ($130) and the most advanced of all, the HTC Vive Pro at $800.
Hardware virtualization
a virtualization of computers or operating systems.It hides the physical characteristics of a computing platform from users, instead showing another abstract computing platform. The software that controls the virtualization used to be called a "control program" but the terms hypervisor or virtual machine monitor are preferred 1. Running one or more applications that are not supported by the host OS 2. Evaluating an alternate operating system 3. Server virtualization 4. Duplicating specific environments 5. Creating a protected environment Virtualization is the consolidation of physical servers into fewer, more powerful, and energy efficient servers. A single physical server can become multiple virtual machines.
VIRTUAL REALITY
an artificial environment which is experienced through sensory stimuli (such as sights and sounds) provided by a computer and in which one's actions partially determine what happens in the environment
Big Data
is high-volume, high-velocity and/or high-variety information assets that demand cost-effective, innovative forms of information processing that enable enhanced insight, decision making, and process automation
Opportunities of Blockchain
● Healthcare ● Voting ● Decentralized IoT ● Properties Rights
New Ways to Run and Operate:Technology Shift from C/S to Cloud Computing
*Bandwidth has increased, and costs for high speed connectivity has declined *Improvements in network storage systems with low latency accessibility *Lower costs, and improved performance for commodity servers *Advances in grid computing and database, OLAP and OLTP switch *Consumerization of enterprise software and demand for systems of engagement
Requirements for Cloud- How does SaaS Provide Value through Non-Functional Qualities ?
*Reliability: 99% availability, conformity with all SLAs and policies, covering all requirements and characteristics that were predefined *Scalability: asks whether the service can be scaled to match the current demand → vertical and horizontal agility to cover and support growth of companies *Availability: you don't have the service on premise, but you want it to be 100% available when you need it, all functionalities are always available, also during maintenance, disaster recovery *Usability: modern look and feel, responsive design, easy to use, access on mobile devices (most SaaS come from the consumerization of customer SW into business) *Data security: requires high integrity, outsourcing requires trust, standards, this is a shared service, so data is stored on shared servers with other tenants
THE FUTURE
The Infinite Display Immersive Advertising and Technology Microsoft Mixed Reality
Possible Societal Benefits
-AI and Machine Learning could have huge economic benefits -Artificial Intelligence, subject to proper regulation and management could create many more jobs -AI could be used to expand further into industries like space exploration -AI could make the world safer -Professional crime agencies are looking to use predictive analysis/machine learning to determine the likelihood of a criminal being a repeat offender -Self driving cars are projected to decrease traffic deaths by 90% -AI could improve healthcare and make it more affordable -Intelligent Sensors are being used to more accurately detect cancerArtificial Intelligence could be utilized to find a cure for diseases like Parkinson's
PaaS: Platform as a Service
-Provides a platform for development and deployment of cloud-based apps -Accessible from web browsers, no need for maintenance of HW / SW -Different subscription models allow different features for the development -Operates below SaaS to provide environment for SaaS development -PaaS is built on virtualization technology, allowing high scalability -Provider takes care of underlying server hardware and infrastructure
The AGL Example
AGL is a company that utilized IoT and the Azure Cloud to build out an solution for veterinarians and pet owners. 1.Collect data from dogs using sensors 2.Have people annotate and organize the data 3.Pass data into GA Tech machine learning program 4.Repeat while monitoring result to ensure quality 5.Pass categorized activities to the database
Possible Organizational Benefits
AI and Machine Learning can save organizations time and money. -AI and Machine Learning reduce labor costs -Machine Learning makes the development of new products/services faster -More reliable data delivered faster allows for better management AI and Machine Learning allow for easier entry to new industries -Machine Learning can give new organizations more power to bring their ideas to market
3D PRINTING IN HEALTHCARE
Bioprinting Ears: takes between 4-6 hours to make Potential for organ transplants (kidneys) Skin printing for injured soldiers Titanium hip implants Cheap prosthetic devices Ear is printed with bioink - a mixture of biodegradable gel and actual human cells Using the patient's own cells: avoid the major problem of rejection
Opportunities: Decentralized IoT
Decentralized networks can reduce infrastructure and maintenance costs Autonomous Decentralized Peer-to-Peer Telemetry (ADEPT) proof-of-concept (PoC) "Economy of Things"
Effects of New Information on Business
IoT enables new ways of obtaining data that allows businesses to gather data and analyze it so that the business owners can make informed decisions.
Iot Devices
Smart Homes Connecting your phone to control your house Wearables Smart Shoes and Smart Watches Healthcare Smart Pills and Smart Dust Monitoring Systems Cars Driverless Cars Sensors to check your car Moving Forward Gartner states that the average household has 9 IoT devices, but that will increase to 500 by 2020
Facebook scandal
The Facebook data privacy scandal centers around the collection of personally identifiable information of up to 87 million people Cambridge Analytica. Cambridge and other companies were able to gain access to personal data of Facebook users due to a variety of factors.
Logic Theorist
The Logic Theorist was a program designed to mimic the problem solving skills of a human It's considered by many to be the first artificial intelligence program and was presented at the Dartmouth Summer Research Project on Artificial Intelligence
Predictive Analytics - Is it Worth It?
The bottom line is, predictive analytics provides... -A way to identify new revenue opportunities -Better customer service -A deeper understanding of customers -Hidden insights -A solid basis to formulating potential hypotheses for business solutions
Total Cost of Ownership (TCO)
Total cost of ownership, or TCO, is a formula that assesses costs and benefits related to the purchase of any IT component. The goal is to calculate a final figure that will reflect the true purchase price,. This is a total cost of ownership comparison per thousand dollars between using software on enterprise premises or software as a service
PROS
Training and planning Safer and more efficient training methods Makes uses of existing tech people own Purpose-built tech becoming cheaper Provides opportunities for increased effectiveness and efficiency in the workplace
History of Blockchain
● Bitcoin is not the foundation of Blockchain ● 1991 → The idea originated ● 1994 → Package Tracking ● 2008 → "Satoshi Nakamoto" - Bitcoin white paper ● 2014 → Ethereum Project = second public blockchain