The age of Artificial Intelligence: An exciting time to be alive!

Transcript

1. The Age of Artificial Intelligence An exciting time to be alive ! By: Tejumade Afonja
2. 2 • In the world today, population is being increased on a daily basis, so also the work that has to be done. • This has led to the increase of manpower in organizations, thereby reducing efficiency. • We can use Artificial Intelligence technologies to automate these routine labor, understand speech or image, make diagnosis in medicine thereby saving manpower as well as increasing productivity and efficiency. Artificial Intelligence - Why?
3. 3 “ Artificial Intelligence can be defined as the science and engineering of making computers behave in a way that, until recently, we thought required human intelligence ” - Andrew Moore
4. 4 The AI Universe
5. 5 Machine Learning is an Artificial Intelligence technology which is simply the science of getting the computer to learn without explicit programming. Machine Learning allows us to extract knowledge from data to form a prediction What is Machine Learning? 1 2
6. Some Practical Applications of Machine Learning 6 Face Recognition • Artificial Intelligence can be used to build facial recognition system which are a viable option for authentication as well as identification. Chatbots • Many organizations embed AI chatbots on their websites to provide online customer care service/assistance.
7. 7 Weather forecast • CNN uses artificial intelligence to predict its daily weather conditions. Diabetic Retinopathy • Artificial intelligence is also widely in healthcare. An example is predicting the different levels of diabetic retinopathy, an eye disease using images of the retina. Some Practical Applications of Machine Learning
8. 8 Linear Algebra Linear Algebra is a branch of Mathematics that concerns systems of linear equations and their representations through matrices and vector spaces Numerical Computation This refers to algorithms that solve mathematical problems by methods that update estimates of the solution via an iterative process Probability & Statistics Probability is the measure of the likelihood that an event will occur in a Random Experiment. Essential Math for Machine Learning
9. 9 Why Linear Algebra? • Dataset and Data Files representation • Image Processing • Image Classification • One - Hot Encoding • Natural Language Processing • Linear Regression • Recommenders Systems • Deep Learning • etc A B C D 1 0 0 1 1 0 1 1 0 1 0 0
10. 10 Essential Linear Algebra Concepts to Understand • Scalar, Vector, Matrices and Tensors • Multiplying Matrices and vectors • Identity and Inverse Matrices • The Determinant • Linear Dependence and Span • Norms • Diagonal and Orthogonal Matrices • Eigen Decomposition • Singular Value Decomposition • Principal Component Analysis
11. 11 Why Probability & Statistics? Probability theory is a mathematical framework for representing uncertain statements. Therefore; • The laws of probability tells us how AI systems should reason, so we design our algorithms to compute or approximate various expressions derived using probability theory. • We use probability and statistics to theoretically analyze the behaviour of proposed AI Systems
12. 12 Why Probability & Statistics? For example, • In supervised Machine Learning, our goal is to learn from labeled data. Data being labeled means that for some inputs X, we know the desired outputs Y • Some possible tasks are: - Identify what’s in an image - Predict the price of a stock given some features about the company - Detect if a file is malicious etc
13. 13 Why Probability & Statistics? • Grouping similar data points together (clustering) • Taking high dimensional data and projecting it into a meaningful lower dimensional space • Representing data with a distribution • Reinforcement Learning; agents taking actions in their environment and observing reward signals based on their behavior.
14. 14 Essential Probability & Statistics to Understand • Random Variables • Discrete Distributions & Continuous Distributions • Joint Probability Distributions • Marginal Probability Distributions • Conditional Probability Distributions • Bayes’ Rule • Independence and Conditional Dependence • Expectation • Variance and Covariance
15. 15 Why Numerical Computation? • Common operation includes optimization i.e finding the value of an argument that minimizes or maximizes a function • Evaluating a mathematical function on a digital computer can be difficult when the functions involves real numbers, which cannot be represented precisely using a finite amount of memory • The goal of Machine Learning model is to reduce the error that a certain prediction makes to get the highest possible accuracy from the model.
16. 16 Essential Calculus to Understand • Overflow and Underflow • Poor Conditioning • Gradient-Based Optimization • Beyond the Gradient: Jacobian and Hessian Matrices • Constrained Optimization
17. 17 Machine Learning in Business • AI and machine learning is capable of automating business intelligence and analytics processes, thereby providing end-to-end solutions. Fraud Detection • Companies like Teradata is an AI firm that sells fraud detection solution to banks. Their platform uses Machine Learning to enhance banking fraud detection. Online Customer Support • Companies like DigitalGenius is an AI-powered customer service tool that uses machine learning and natural language processing to completely automate the consumer support process
18. Healthcare Benefits • Companies like Ubenwa is saving newborn lives by enabling quick and cost-effective diagnosis of birth asphyxia from infant cry using machine learning system that can take as input the infant cry, analyse the amplitude and frequency patterns in the cry 18 Machine Learning in Business Cybersecurity Defense • Companies like DarkTrace are focused on using unsupervised Machine Learning to analyze network data at scale thereby responding to cyber threat somewhere in the world every 3 seconds
20. 20 Using Machine Learning in Business: Using AI in your Company 1. Examine how your competitors are using AI e.g How are Hospitals using AI? How is Google using AI? etc 2. Decide what AI can do for your business 3. Search for and compare vendors 4. Implement an AI Project
21. 21 Using Machine Learning in Business: Approaches that realizes business value 1. Business Problem (Definition, value, stakeholders, priority, investment) - start small and narrow until you realize what the business value is 2. Data (Availability, provenance, security, coverage, cleaning, augmentation, annotation, refreshing, pipeline development) 3. Model Building (Feature extraction, hyperparameters, tuning, selection, benchmarking) 4. Deploy & Measure (Business value measurement, AB testing, versioning, business process integration) 5. Active Learning & Tuning (Bias mitigation, ground truth & success monitoring, version control) 6. Rinse & Repeat
22. 22 AI AS A SERVICE - Instadeep is a leading AI company in Africa. InstaDeep builds decision making systems for mobility, logistics using the latest AI breakthroughs and our own in-house innovation while also providing training and access to some of the best professionals in their field, allowing future AI leaders and experts the opportunity to advance in line with the rapidly developing industry, ensuring talent retention and full in-house expertise. Karim Beguir Co-founder & CEO
23. 23 AI Saturdays Lagos AI6 is an active learning community which promotes artificial intelligence in Lagos by organizing structured groups around core AI fields like Machine Learning, Computer Vision, and Natural Language Processing. Our goal is to democratize AI by creating a community to help enable studying, researching and building AI products for our ecosystem.
24. 24 Incase you were looking to learn about how to get started with Machine Learning, check out this codelab and see if you like it :) https://www.tensorflow.org/beta/tutorials/keras/basic_regression Also, join AI Saturdays Lagos & TensorFlow Lagos Community :-p Extras
25. 25 Get in Touch linkedin.com/company/instadeep twitter.com/instadeepai, twitter.com/aisaturdaylagos, twitter.com/tensorflowlagos
26. 26 1. https://www.deeplearningbook.org/ 2. https://towardsdatascience.com/calculus-in-data-science-and-its-uses-3f3e1b5e5b35 3. https://towardsdatascience.com/probability-fundamentals-of-machine-learning-part-1-a156b 4703e69 4. https://towardsdatascience.com/machine-learning-probability-statistics-f830f8c09326 5. https://www.sas.com/en_gb/insights/articles/analytics/applications-of-artificial-intelligence. html 6. https://www.forbes.com/sites/forbestechcouncil/2018/09/27/15-business-applications-for-a rtificial-intelligence-and-machine-learning/#554b88a3579f 7. https://www.forbes.com/sites/julianmitchell/2018/09/05/how-ai-machine-learning-and-other -disruptive-trends-are-defining-the-future-of-customer-service/#331988fe4cdf 8. https://emerj.com/ai-sector-overviews/artificial-intelligence-fraud-banking/ 9. https://www.datamation.com/artificial-intelligence/artificial-intelligence-in-business.html References
27. 27 Thank You © Copyright 2019 InstaDeep.com. All Rights Reserved.