Ai digital (without videos)

1. Using Artificial Intelligence & Machine Learning to Transform Digital Experiences AI Everywhere & Nowhere GLOBAL ARTIFICIAL INTELLIGENCE LEAD Dr. Anand S. Rao www.pwc.com

2. PwC New Services and Emerging Technology – AI Lab AI: Computer system or agent that can sense, think, and act in an environment to achieve a purpose 2 AI that can sense… Hear See Speak Feel AI that can think… Understand Reason PlanLearn AI that can act… Physical Sensors Digital Effectors • Knowledge Rep. • Reasoning • Machine Learning • Deep Learning • Simulation • Robotic process automation • Deep question & answering • Collaborative systems • Adaptive systems • Natural language • Audio & speech • Machine vision • Navigation • Visualization Statistics Econometrics Optimization Complexity Theory Computer Science Game Theory FOUNDATION LAYER

3. PwC New Services and Emerging Technology – AI Lab Today’s discussion AI and Digital Experiences From Consumer to Enterprise Digital Experiences Opportunities, Risks, and Implications for Enterprises 01 02 03 3

4. PwC New Services and Emerging Technology – AI Lab AI and Digital Experiences 4 01

5. PwC New Services and Emerging Technology – AI Lab Sizing the Prize: AI in productivity & consumption gains 5 Are you ready to exploit the opportunities from AI & overcome the challenges? Global GDP Impact of AI through 2030 GlobalGDPupliftduetoAI ($intrillions) 2030 IMPACT: $15.7T Consumption Contribution: 60% Source: PwC Analysis; Productivity Contribution: 40%

6. PwC New Services and Emerging Technology – AI Lab AI and Digital Experience in the Consumer World 6 AI as UI (Ubiquitous Intelligence) AI is being embedded in devices, things, people and is becoming ubiquitous in our daily life AI as No UI (User Interface) Conversational, chat, haptic and brain-machine interfaces will augment existing interfaces AI as AAAAI AI is being used as automated, assisted, augmented, and autonomous intelligence

7. PwC New Services and Emerging Tech Confidential information for the sole benefit and use of PwC’s client. 7 Google Search: Auto-completion (N-gram) Google Duplex: Automating reservations Four Uses of AI: Ai is moving beyond automating and assisting humans in hardwired situations to becoming more adaptive – augmenting and becoming more autonomous No human in the loopHuman in the loop Hardwired / specific systems Adaptive systems Automated Intelligence 1 Assisted Intelligence 2 Augmented Intelligence 3 Autonomous Intelligence 4 + AI as AAAAI (Automated-Assisted-Augmented-Autonomous Intelligence)

8. PwC New Services and Emerging Technology – AI Lab From Consumer to Enterprise Digital Experiences 8 02

9. PwC New Services and Emerging Technology – AI Lab Companies are starting their AI investments in automation, with long- term thinkers also exploiting augmented/autonomous AI • AI techniques enhance the efficiency of activities across the business value chain, but machines do not dynamically adapt to changing data Automated Intelligence Assisted Intelligence Augmented Intelligence Autonomous Intelligence Degree of Advancement High risk - Big bets, transforming business models Low risk - Quick wins happening right now • Computational algorithms begin to adapt to changing data; machines do not automatically make decisions, however they put humans in the best place to make decisions • AI techniques used by businesses to automate the decision making process with the absence of human intervention • Automation of repetitive tasks that include both manual and cognitive aspects

10. Fukoku Mutual insurance company is automating business processes to reduce labor Automation Assisted Intelligence Augmented Intelligence Autonomous Intelligence Japanese white-collar workers are already being replaced by artificial intelligence Robotic Process Automation Natural-Language Processing +30% increase in labor productivity 110-140M reduction in workers by 2025 3x benefit over offshoring Robotic Process Automation (RPA) Capabilities RPA vendor solutions are dominating the market for automating processes but have limitations on the extent and scope of impact they can have $2M annual savings *at $150K maintenance

11. Construction company used drones and deep learning to monitor construction site progress and track assets Automation Assisted Intelligence Augmented Intelligence Autonomous Intelligence Drones and artificial intelligence can empower complex infrastructure projects Computer Vision Machine Learning Deep Learning 60% reduction in operational cost $3M annual savings on a $5M project 700 hours of labor savings on one project From a drone aerial picture company was able to produce segmented output of different objects in that image Key: Background Trees Asphalt ConcreteCars Reinforcement

12. A Global Pharmaceutical company used NLP to extract adverse drug interaction from multiple unstructured data sources Automation Assisted Intelligence Augmented Intelligence Autonomous Intelligence NLP architecture and pipeline are critical to automate cognitive processes and generate insights Natural-Language Processing Machine Learning Deep Learning +20% Annual growth of adverse events $14-18 M annual savings on current base 96% diagnostic accuracy Clinician notes Social media Medical literature Tokenization Grammar Parsing Text Normalization Text Cleaning Word Disambiguation Vectorization SourcesProcess 1 Gathering key information output, e.g., patient sneezes (event) 2 Deep Learning of Latent Relationships, e.g., sneezing and antihistamine 3 35%-45% Savings in processing costs

13. Global airline used predictive aircraft maintenance to reduce maintenance related costs from Delays & Cancellations Automation Assisted Intelligence Augmented Intelligence Autonomous Intelligence Aircraft predictive maintenance Natural-Language Processing Machine Learning 15% reduction in delays due to maintenance -$25M Cost reduction Provides airline clients a deep analysis on aircraft fault messages and text analytics on maintenance logs to find significant signals that cause delay and cancellation events Diagnostic Enables reliability engineers to monitor fleet health and identify trends, chronic aircraft and parts Fleet Reliability Provides maintenance controllers indication of potential failures at the aircraft component level that necessarily result in delays and cancellations (D&Cs) 2-5 days out enabling maintenance intervention Alerting Airline Predictive Maintenance Solution +0.9% on time performance

14. Global auto manufacturer gamified its strategy to evaluate go-to-market scenarios for a new rideshare and autonomous vehicle business Automation Assisted Intelligence Augmented Intelligence Autonomous Intelligence Simulations to evaluate go-to-market scenarios Machine Learning Deep Learning 200,000 go-to-market scenarios evaluated 170 M Miles delivered by 10,000 vehicles Simulations Select Cities Strategies Random Seeds Market Condition Used demographic models and demand estimator • Price • Aggressiveness • Marketing • Customer serviceEach strategy repeated 10 times to account for randomness Different conditions of consumer acceptance Digital Twins of consumers… Socio- demographics …were modelled under a set of scenarios Transport choice (commute, errand, weekend) City topology $1+ bn Acquisition of AV technology start-up

15. Leading companies are moving from descriptive and diagnostic analytics to prescriptive and cognitive analytics where AI plays a greater role Machine Learning Deep Learning Describe, summarize and analyze historical data Recommend ‘right’ or optimal actions or decisions Monitor, decide, and act autonomously or semi-autonomously Predict future outcomes based on facts from the past and simulations Descriptive Predictive Prescriptive Cognitive IncreasingBusinessValue Identify causes of trends and outcomes Diagnostic Increasing Sophistication of Data & Analytics (What happened?) (Why it happened?) (What could happen?) (What should be done?) (How do we adapt to change?)

16. PwC New Services and Emerging Tech How do enterprises move from a ‘mobile first’ to an ‘AI first’ mindset? 16PwC New Services and Emerging Technology – AI Lab

17. PwC New Services and Emerging Technology – AI Lab Opportunities, Risks and Implications for Enterprises 17 03

18. PwC 18 As a result, we focus on addressing the challenges that organizations face, as they seek to exploit this new technology in their enterprises Data, Data, Data Everywhere – but not the right kind for AI Collecting, organizing, storing, safeguarding, labelling and exploiting the data for enterprise applications Making AI more human AI that is more ‘human’, that can interact with humans at the right level, learn from humans, teach humans, and resolve ethical dilemmas Acquiring, developing, and retaining the right talent to explore and exploit AI Ensuring the right mix of business domain expertise, computing experience, statistical and mathematical knowledge in teams Building safe & robust AI that is trustworthy Building AI that can explain itself, is transparent, can be controlled, and is without bias. Increased vulnerability and disruption to business 77% Potential for biases and lack of transparency 76% Ensuring governance and rules to control AI 73% Risk to stakeholders’ trust and moral dilemmas 71% Potential to disrupt society 67% Lack of adequate regulation 64% Starting point – From data, automation, or analytics Aligning AI initiatives across the enterprise emanating from big data, analytics, and automation initiatives What’s holding AI back in the enterprise? Source: PwC CEO Pulse Survey, 2017 Q: Which of the following issues surrounding AI adoption concern you the most Base: 239

19. PwC New Services and Emerging Technology – AI Lab Benefiting from AI requires separating myths from facts 19 Myth 1: Artificial Intelligence is a distinct monolithic area of study Fact 1: Artificial Intelligence is an interdisciplinary area with many distinct sub-fields Myth 2: All types of problems can be solved by a single AI solution (e.g., ….insert your favorite solution) Fact 2: Different types of problems require different type of AI techniques and solutions to be used Myth 3: Machine Learning automatically (magically) learns from data without any human intervention Fact 3: Machine Learning requires a laborious process of acquiring and cleansing large amounts of data, selecting, training, and guiding the algorithm Searching, Querying & Conversing Describing, Classifying, Understanding & Visualizing Diagnosing, Discovering & Reasoning Trending, Forecasting, Projecting & Predicting Simulating, Learning, Optimizing, & Adapting Recognizing, Sensing, and Recommending AI Uses

20. PwC New Services and Emerging Technology – AI Lab Start from the business value chain and metrics to be improved to get better ROI from AI 20 Operations & Development Product Development Service & Support Operations Outbound Logistics Sales & Distribution Customers & Marketing Strategy & Growth Supply Chain & Procurement Finance, HR, Planning Inbound Logistics How will we ensure our product supply is meeting demand? VP, Supply Chain How can we engage with our customers to enhance their experience? Director, Marketing How can we grow our market share and which markets to enter, exit or expand? Director, Strategy How do we innovate and introduce new products and services? Director, Products How do we increase customer satisfaction and retain more customers? Director, Service How can we reach more customers and price our products to increase sales? Director, Sales How can we increase efficiency and effectiveness of our operations? Director, Operations How can we get a better return on our talent, capital, and assets? Director, Finance & HR • Market Share • Customer Experience • Acquisition Rate • Innovation Rate • Operational Efficiency • Customer Satisfaction • Talent Retention • Inventory Turn Over 300+ AI Use Cases Across 8 Sectors – Sizing the Prize

21. PwC New Services and Emerging Technology – AI Lab Focus on a few key areas of AI to get traction and critical mass of expertise 21 Data Eng./Model Ops Automated ML Simulation & RL Responsible AIEmbodied AI ▪ Natural Language processing and text mining ▪ Natural Language generation ▪ Chatbots and discourse understanding ▪ Sentiment & emotion analysis ▪ Speech-to-text and text-to- speech ▪ Convolutional Neural Nets ▪ Recursive Neural Nets ▪ Capsule Networks ▪ Generative Adversarial Networks ▪ Deep reinforcement learning ▪ Hybrid learning models ▪ Regression & classification ▪ Bayesian learning ▪ Probabilistic programming ▪ Anomaly detection ▪ Optimization techniques ▪ Support Vector Machines ▪ Various supervised, semi- supervised, and unsupervised techniques ▪ Big data architecture ▪ Big and Fast data ▪ Apache tools ▪ Cloud computing ▪ Cloud ML – AWS, GCP, Azure ▪ Machine Learning deployment ▪ Agent-based simulation ▪ Reinforcement learning ▪ Augmented and synthetic data generation ▪ System dynamics modeling ▪ ’Digital Twins’ ▪ Calibration of models ▪ IoT and Industrial IoT – Edge computing and Smart sensors ▪ Drone – Autonomy & Image analytics ▪ Robots – Navigation & Learning ▪ Brain-Machine Interfaces ▪ Explainable AI ▪ Beneficial AI ▪ ‘Black box’ Interpretability ▪ Maturity models ▪ Ethics and Law ▪ AI Governance ▪ AI Controls framework ▪ Automated data preparation ▪ Automated feature engineering ▪ Automated algorithm selection ▪ Automated explanation generation ▪ Meta-model inference Natural Language Machine Learning Deep Learning

22. PwC 22 Balance the opportunities with the significant risks that need to be assessed, mitigated and managed Control • Risk of AI going ‘rogue’ (e.g., Tay Chatbot) • Inability to control malevolent AI • Swarm drones Security • Cyber intrusion risks • Privacy risks • Open source software risks • Digital, Physical, Political security Societal • Risk of Autonomous Weapons proliferation • Risk of ‘intelligence divide’ Ethical • ‘Lack of Values’ risk • Value Alignment risk • Goal Alignment risk Economic • Job displacement risks • ‘Winner-takes-all’ concentration of power risk • Liability risk Performance • Risk of Errors • Risk of Bias • Risk of Opaqueness • Risk of stability of performance • Lack of feedback process Risk Robust & Safe AI Beneficial AI Responsible AI

23. PwC New Services and Emerging Technology – AI Lab 23 Start from business decisions 01 Demonstrate value through pilots before scaling 02 Blend intuition and data-driven insights 03 Fail forward – test and learn culture 05 Focus on Responsible AI from the start 06 Address ‘big data’ – don’t forget ‘lean’ data 04 Six success factors to derive maximum benefits from artificial intelligence

24. PwC New Services and Emerging Technology – AI LabPwC’s Digital Services Thank you. © 2018 PwC. All rights reserved. Not for further distribution without the permission of PwC. “PwC” refers to the network of member firms of PricewaterhouseCoopers International Limited (PwCIL), or, as the context requires, individual member firms of the PwC network. Each member firm is a separate legal entity and does not act as agent of PwCIL or any other member firm. PwCIL does not provide any services to clients. PwCIL is not responsible or liable for the acts or omissions of any of its member firms nor can it control the exercise of their professional judgment or bind them in any way. No member firm is responsible or liable for the acts or omissions of any other member firm nor can it control the exercise of another member firm’s professional judgment or bind another member firm or PwCIL in any way. Dr. Anand S. Rao Global AI Lead anand.s.rao@pwc.com @AnandSRao

25. PwC New Services and Emerging Tech Confidential information for the sole benefit and use of PwC’s client. 25 Google Search: Auto-completion (N-gram) Google Duplex: Automating reservations Man-Machine combination to create unique advantage Large teams of people watched and tagged movies and shows (36 page manual) Created 76,897 micro- genres AI as Ubiquitous Intelligence - Netflix

26. PwC New Services and Emerging Tech Confidential information for the sole benefit and use of PwC’s client. 26 Google Search: Auto-completion (N-gram) Google Duplex: Automating reservations Conversational Interfaces: Increasingly conversational interfaces will replace keyboard input, where appropriate AI as No UI – Conversational Interfaces

27. PwC New Services and Emerging Tech Confidential information for the sole benefit and use of PwC’s client. 27 Google Search: Auto-completion (N-gram) Google Duplex: Automating reservations Brain-Machine Interfaces: In our AI Lab we are experimenting with EEG & biometric devices to control devices from brain wave patterns AI as No UI – Brain-Machine Interfaces OpenBCI Mark VI Headset - 16 ChannelsEmpatica E4 Wristband

Ai digital (without videos)

Vous êtes en train de lire un aperçu.

Consultez-les par la suite

Ai digital (without videos)

Plus De Contenu Connexe

Diaporamas pour vous (20)

Similaire à Ai digital (without videos) (20)

Plus récents (20)