Demystifying AI & Machine Learning

In this presentation, I will provide a brief background of the origins and history of AI and Machine Learning, talk about the steps involved in creating Machine Learning models and then finish with how your business can set themselves up for success.

Transcript

1. Demystifying artificial intelligence Ricardo Sueiras Evangelist Amazon Web Services 100% Babel Fish Free @094459 ricsue@amazon.com
2. Myth #1 - AI is the flavour of the month
3. Henrietta Swan Leavitt She discovered a relationship between the absolute brightness of the star and the period of pulsation so by observing their period, the absolute brightness could be inferred. July 4, 1868 – December 12, 1921
4. Abraham Wald Statistical Research Group (SRG) Amazon Confidential
5. John McCarthy (1927-2011) 1956 - Coined the term “Artificial Intelligence” Fact #1 – The term ”AI” is 60 years old “every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.” - A PROPOSAL FOR THE DARTMOUTH SUMMER RESEARCH PROJECT ON ARTIFICIAL INTELLIGENCE
6. The goals of AI have evolved Emulate Humans Technology at the Centre Augment Human capabilities People at the Centre 1950’s Today Alan Turing (1912-1954) Turing Test
7. aka « You’re not smart enough » Myth #2 - AI is dark magic
8. A bit of Science, a lot of Engineering Fact #2 - AI is math, code and chips Amazon Confidential
9. March, 2017 100 School girls aged 14-16 split into teams of 8 build an AI enabled device in four hours at the Science Museum. Every team builds a working device within the time.
10. Myth #3 - AI is not production-ready
11. *Total economic impact of AI in the period to 2030 according to PwC $15.7 Trillion Game Changer* Fact #3 –”AI” is ready for business +26% GDP boost
12. Amazon Confidential Algorithms Data Accelerate CPU/GPU Cloud Why now?
13. Classical Programming Machine Learning Data Rules Data Answers Answers Rules What is machine learning?
14. A process by which a computer system makes decisions based on rules that it learned on its own (through the use of data) What is machine learning?
15. What is AI and Machine Learning? Stuff computers are good at Stuff humans are good at Biological Brains Creativity, Emotion, Intuition and Empathy AI The Future General AI Performing specific tasks that mimic human actions. Process and respond quickly, learn from huge data sets Machine Learning Computer Models that adjust themselves based on data (inputs) A Subset of ML which makes computation of multi-layer neural networks feasible Deep Learning Intelligent Systems Amazon Confidential
16. Machine Learning Artificial Intelligence Deep Learning “ability to make decisions” “ability to learn rules” “ability to learn concepts” Broad Niche Machine Learning in the broader context
17. Step 1: Problem Statement Step 2: Creating and Training a Model Step 3: Inference What is machine learning?
18. Step 1: PROBLEM What am I trying to solve for? How do I get started?
19. Supervised vs. Unsupervised vs Reinforcement • Labelled training data • Want to predict labels of new, unlabeled data • Ex: Classification, K-nearest neighbor • Algorithm finds trends in data, optimization is algorithm-reliant • Ex: Automated clustering, k-means, data exploration • Complex problem/reward space • Consecutive actions by agent result in an outcome/score • Ex: Agent for autonomous driving Three classes of problem
20. Types of machine learning Machine Learning Unsupervised Learning Supervised Learning Reinforcement Learning No labeled data - Find the hidden structure – classification, categorization, etc Labeled data – predict an outcome or future with greater precision Agent based – reward engine, learn via experience Spoiler! You will be diving deep during your AI Academ y Customer Segmentation Targetted Marketing Recommended Systems Big Data Visualisations Structure discovery Forecasting Predictions Diagnostics / Anomaly detection Image Classification Fraud Detection New Insights Robot Navigation Skill Acquisition Game AI Learning Tasks Real Time Decisions
21. Step 2: TRAINING How do I create and train a model?
22. What is ML: Training Data Training Data: Class 1 Training Data: Class 2
23. Model Decision Boundary What is ML: Model
24. The algorithm is not visible in this diagram, because is not an object or property, it is a process. ? Each algorithm behaves differently, and has its own tradeoffs (no objectively best algorithm!) What is ML: Algorithm
25. Step 3: INFERENCE How do I use a model?
26. What is ML: Decision Area Model Decision Boundary Decision Area: Class 1 Decision Area: Class 2
27. Business Problem defines your Optimization Parameter Step 1: PROBLEM Step 2: TRAINING Step 3: INFERENCE Determines Gets wrapped as function Machine Learning, in three steps
28. Reinforcement Learning
29. Reinforcement learning in the real world
30. Setting your organization up for AI and ML success Putting it all together
31. Predictive Analytics Real time Analytics Amazon Confidential B.I Eventually everything connects: IoT, Mobile, API’s & Applications Large amounts of data processed in real time Using data to build machine learning models Building on data science • Big Data and Data Lakes • Data Cleansing • Data Preparation • Analysis Machine Learning requires extensive (BIG) data services Data Science provide the foundation
32. Amazon S3 | AWS Glue AWS Direct Connect AWS Snowball AWS Snowmobile AWS Database Migration Service AWS IoT Core Amazon Kinesis Data Firehose Amazon Kinesis Data Streams Amazon Kinesis Video Streams On-premises Data Movement Amazon SageMaker AWS Deep Learning AMIs Amazon Rekognition Amazon Lex AWS DeepLens Amazon Comprehend Amazon Translate Amazon Transcribe Amazon Polly Amazon Athena Amazon EMR Amazon Redshift Amazon Elasticsearch Service Amazon Kinesis Amazon QuickSight Analytics Machine Learning Real-time Data Movement Data Lake Before you have a machine learning strategy, you need a data strategy
33. What is the answer the question of ‘Life, the Universe, and Everything.’ “You know nothing of future time,” pronounced Deep Thought, “and yet in my teeming circuitry I can navigate infinite delta streams of future probability and see that there must one day come a computer whose merest operational parameters I am not worthy to calculate, but which it will be my eventual fate to design.” Amazon Confidential
34. Start with your problem/question • Start with a defined and concrete business problem you want to solve • Problem definition needs to be well scoped and S.M.A.R.T • Netflix – how do we help customers better find films that we think they will like based on what they are interested in? Value Of Increase in data and technology Questions – are you asking good questions? Answers Amazon Confidential
35. Latency Cost Scalability Customizability Ideal solution considers the intersection of
36. Employee Skillset Agility Requirements Foundational Infrastructure Organisation readiness
37. Our mission at AWS Put machine learning in the hands of every developer and data scientist
38. © 2019, Amazon Web Services, Inc. or its Affiliates. FRAMEWORKS INTERFACES INFRASTRUCTURE AI Services Broadest and deepest set of capabilities THE AWS ML STACK VISION SPEECH LANGUAGE CHATBOTS FORECASTING RECOMMENDATIONS ML Services ML Frameworks + Infrastructure P O L L Y T R A N S C R I B E T R A N S L A T E C O M P R E H E N D & C O M P R E H E N D M E D I C A L L E X F O R E C A S T R E K O G N I T I O N I M A G E R E K O G N I T I O N V I D E O T E X T R A C T P E R S O N A L I Z E Ground Truth Notebooks Algorithms + Marketplace Reinforcement Learning Training Optimization Deployment Hosting Amazon SageMaker F P G A S E C 2 P 3 & P 3 D N E C 2 G 4 E C 2 C 5 I N F E R E N T I A G R E E N G R A S S E L A S T I C I N F E R E N C E D L C O N T A I N E R S & A M I s E L A S T I C K U B E R N E T E S S E R V I C E E L A S T I C C O N T A I N E R S E R V I C E
39. 1 Create the loop Connect technology initiatives with business outcomes 2 Assess your structured and unstructured data sources Advance your data strategy ? 3 Put machine learning in the hands of your developers Organize for success Set yourself up for success