Week 1 Introduction to AI.pdf

INTRODUCTION TO ARTIFICIAL INTELLIGENCE, MACHINE LEARNING, AND INTELLIGENT AGENTS Presenter: Dr. Rami Hawileh American University of Sharjah Prepared by Dr. Salam Dh...

INTRODUCTION TO ARTIFICIAL INTELLIGENCE, MACHINE LEARNING, AND INTELLIGENT AGENTS Presenter: Dr. Rami Hawileh American University of Sharjah Prepared by Dr. Salam Dhou, CSE Some content are from slides shared in: https://www-users.cse.umn.edu/~kumar001/dmbook/index.php 1 Table of Content 2  Importance of AI and Data Mining  Artificial Intelligence  Machine Learning  Intelligent Agents 2 1  Large-scale Data is Everywhere! 3  There has been enormous data growth in both commercial and scientific databases due to advances in data generation and E-Commerce collection technologies. Cyber Security  New mantra: Gather whatever data you can whenever and wherever possible.  Expectations: Gathered data will Social Networking: Twitter Traffic Patterns have value either for the purpose collected or for a purpose not envisioned. Sensor Networks Computational Simulations 3 Why Data Mining? Commercial Viewpoint 4  Lots of data is being collected and warehoused.  Web data ◼ Yahoo has Peta Bytes of web data ◼ Facebook has billions of active users  Purchases at department/grocery stores, e-commerce ◼ Amazon handles millions of visits/day  Bank/Credit Card transactions  Computers have become cheaper and more powerful.  Competitive Pressure is Strong.  Providebetter, customized services for an edge (e.g. in Customer Relationship Management) 4 2  Why Data Mining? Scientific Viewpoint 5  Data collected and stored at enormous speeds  Remote sensors on a satellite ◼ NASA EOSDIS archives over petabytes of earth science data / year Sky Survey Data  Telescopes scanning the skies ◼ Sky survey data  High-throughput biological data fMRI Data from Brain  Scientific simulations ◼ Terabytes of data generated in a few hours Gene Expression Data  Data mining helps scientists  In automated analysis of massive datasets  In hypothesis formation Surface Temperature of Earth 5 Importance of AI 6  AI has many uses from boosting vaccine development to automating detection of potential fraud.  AI companies raised $66.8 billion in funding in 2022, according to CB Insights research*, more than doubling the amount raised in 2020.  Because of its fast-paced adoption, AI is making waves in a variety of industries.  AI has positive impact on different sectors, such as*:  Better Medicine  Safer Banking  Innovative media  Agriculture  Renewable energy *https://www.cbinsights.com/research/report/ai-trends-2022/ 6 3  Importance of AI 7  Better Medicine  A 2021 World Health Organization report* noted that integrating AI into the healthcare could lead to benefits such as more informed health policy and improvements in the accuracy of diagnosing patients.  Some applications of AI in healthcare  Improving medical diagnosis  Speeding up drug discovery  Transforming patient experience  Performing robotic surgery Improving health care and reducing costs *https://www.who.int/publications/i/item/9789240029200 7 Importance of AI 8  Better Medicine: Improving medical diagnosis  Every year, roughly 400,000 hospitalized patients suffer preventable harm, with 100,000 deaths*. In light of that, the promise of improving the diagnostic process is one of AI’s most exciting healthcare applications.  Incomplete medical histories and large caseloads can lead to deadly human errors.  AI can predict and diagnose disease at a faster rate than most medical professionals.  Many healthcare companies are using AI to reduce errors and save lives. *Rodziewicz TL, Houseman B, Hipskind JE. Medical Error Reduction and Prevention. 2023 May 2. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2023 Jan–. PMID: 29763131. 8 4  Importance of AI 9  Better Medicine: Speeding up drug discovery  The drug development industry is affected by the high development costs and research that takes thousands of human hours. Putting each drug through clinical trials costs an estimated average of $1.3 billion*, and only 10 percent of those drugs are successfully brought to market.  Due to breakthroughs in technology, biopharmaceutical companies are leveraging the efficiency, accuracy and knowledge AI can provide.  Several companies use AI to develop the next generation of medicines. *https://www.cbinsights.com/research/clinical- trials-ai-tech-disruption/ 9 Importance of AI 10  Better Medicine: Transforming patient experience  Inthe healthcare industry, time is money. Providing a seamless patient experience allows hospitals, clinics and physicians to treat more patients on a daily basis.  Studies show hospitals whose patients have positive experiences have higher profits*, while negative reviews can lead to financial losses.  AI is being used to help healthcare facilities better manage patient flow. *https://www.medicaleconomics.com/view/the-link- between-financial-success-and-patient-satisfaction 10 5  Importance of AI 11  Better Medicine: Performing robotic surgery  Robot-assisted surgery* is becoming very popular. Robots are being used to help with everything from minimally invasive procedures to open heart surgery.  According to the Mayo Clinic, robots help doctors perform complex procedures with a precision, flexibility and control that goes beyond human capabilities**.  Surgeons can control a robot’s mechanical arms while seated at a computer console as the robot gives the doctor a three dimensional, magnified view of the surgical site that surgeons could not get from relying on their eyes alone. The surgeon then leads other team members who work closely with the robot through the entire operation.  Robot-assisted surgeries have led to fewer surgery-related complications, less pain and a quicker recovery time. *https://builtin.com/robotics/surgical-medical-healthcare-robotics-companies **https://www.mayoclinic.org/tests-procedures/robotic-surgery/about/pac-20394974 11 Importance of AI 12  Safer Banking  Business Insider Intelligence’s 2022 report on AI in banking* found more than half of financial services companies already use AI solutions for risk management and revenue generation.  The application of AI in banking could lead to upwards of $400 billion in savings. *https://www.businessinsider.com/ai-in-banking-report 12 6  Importance of AI 13  Innovative Media  AI has also made its mark on entertainment.  The global market for AI in media and entertainment is estimated to reach $99.48 billion by 2030, growing from a value of $10.87 billion in 2021, according to Grand View Research*.  That expansion includes AI uses like recognizing plagiarism and developing high-definition graphics. *https://www.grandviewresearch.com/industry-analysis/artificial- intelligence-ai-media-entertainment-market-report 13 Importance of AI 14  Great Opportunities to Solve Society’s Major Problems Predicting the impact of climate change Reducing hunger and poverty by increasing agriculture production Finding alternative/ green energy sources 14 7  Artificial Intelligence 15 Artificial Intelligence (AI): Intelligence demonstrated by machines, as opposed to the natural intelligence displayed by animals including humans. AI is a wide-ranging branch of computer science concerned with building smart machines capable of performing tasks that typically require human intelligence. Human Intelligence Artificial Intelligence 15 Machine Learning 16  Machine learning is an application (subset) of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.  In traditional programming, a computer engineer writes a series of directions that instruct a computer how to transform input data into a desired output. Instructions are mostly based on an IF-THEN structure: when certain conditions are met, the program executes a specific action.  Machine learning, on the other hand, is an automated process that enables machines to learn and take actions based on past observations.  Machine Learning is needed for tasks that are too complex for humans to code directly/explicitly (because there are too many variables to consider and too many nuances in the model). 16 8  Artificial Intelligence vs Machine Learning 17  While artificial intelligence and machine learning are often used interchangeably, they are two different concepts:  AI is the broader concept – machines making decisions, learning new skills, and solving problems in a similar way to humans.  Machine learning is a subset of AI that enables intelligent systems to autonomously learn new things from data. Artificial Intelligence Machine Learning Deep Learning 17 Types of Machine Learning 18  There are different machine learning methods and algorithms, which are basically sets of rules that machines use to make decisions.  The five most common and most used types of machine learning are: 1. Supervised Learning 2. Unsupervised Learning 3. Semi-Supervised Learning 4. Reinforcement Learning 5. Deep Learning (DL) 18 9  Supervised Learning 19  Supervised learning models make predictions based on labeled/annotated training data.  Supervised learning is the most common and popular approach to machine learning.  It’s “supervised” because these models need to be fed manually labeled sample data to learn from.  For example, if you want to automatically detect spam, you would need to feed a machine learning algorithm examples of spam emails with their labels (SPAM) and others that are important with their labels (NOT SPAM).  There are two types of supervised learning:  Classification  Regression 19 Supervised Learning 20 Learning Concept Training Phase Results in a trained model Prediction Phase Trained model 20 10  Unsupervised Learning 21  Unsupervised learning algorithms uncover insights and relationships in unlabeled data. In this case, models are fed input data but the desired outcomes/labels are unknown.  The models are not trained with the “right answer,” so they must find patterns on their own. 21 Unsupervised Learning- Clustering 22  Clustering is one of the most common types of unsupervised learning, which consists of grouping samples based on similarity.  Different similarity metrics can be used.  Similarity can be how close the points are to each others in a Euclidean space. 22 11  Clustering: Applications 23  Understanding  Custom profiling (persona) for targeted marketing  Group related documents for browsing  Group genes and proteins that have similar functionality  Group stocks with similar price fluctuations  Summarization  Reduce the size of large data sets Use clustering on the customer database to identify patterns and group customer according to their purchasing behaviors and interests Trying to determine the appropriate audience for a specific product Advertising the product to the targeted audience 23 Unsupervised Learning- Association Rule Mining 24  Association rule mining is another type of unsupervised learning  It produce association rules which are used to predict occurrence of an item based on occurrences of other items.  There are algorithms to extract the association rules from datasets  Given a dataset of transactions: TID Items 1 Bread, Milk 2 Bread, Diaper, Water, Eggs 3 Milk, Diaper, Water, Coke 4 Bread, Milk, Diaper, Water 5 Bread, Milk, Diaper, Coke One of the induced rules is: {Diaper} → {Milk} It means customers who bought Diaper, they also bought Milk. Order of the rule matters! The converse is not always true. 24 12  Association Rule Mining: Applications 25  Market-basket analysis  Associationrule mining are used for sales promotion, shelf management, and inventory management  For example, shops place associated items next to each others so customers can see and buy, which will consequently boost the shops’ sales! 25 Association Rule Mining: Applications 26  Medical Informatics  Associationrule mining are used in the medical domain to find combination of patient symptoms and test results associated with certain diseases. 26 13  Semi-Supervised Learning 27  In semi-supervised learning, training data is split into two portions:  A small amount of labeled data  A larger set of unlabeled data  The goal of a semi-supervised learning model is to make effective use of all of the available data, not just the labelled data like in supervised learning.  Making effective use of unlabelled data may require the use of or inspiration from unsupervised methods such as clustering and density estimation to find groups or patterns in the data.  Once groups or patterns are discovered, supervised methods or ideas from supervised learning may be used to label the unlabeled examples. https://datawhatnow.com/pseudo-labeling- semi-supervised-learning/ 27 Semi-Supervised Learning 28  Semi-supervised learning is gaining popularity, especially for tasks involving large datasets such as image classification.  Semi-supervised learning doesn’t require a large number of labeled data, so it’s faster to set up, more cost-effective than supervised learning methods, and ideal for businesses that receive huge amounts of data. 28 14  Reinforcement Learning 29  Reinforcement learning is concerned with how a software agent (or computer program) ought to act in a situation to maximize the reward.  Reinforced machine learning models attempt to determine the best possible path they should take in a given situation. They do this through trial and error. Since there is no training data, machines learn from their own mistakes and choose the actions that lead to the best solution or maximum reward.  This machine learning method is mostly used in robotics and gaming. 29 Deep Learning 30  Deep learning is a subset of machine learning  It is based on Artificial Neural Networks (ANN), a type of computer system that emulates the way the human brain works.  Deep learning algorithms are built with multiple layers of interconnected neurons, allowing multiple systems to work together simultaneously, and step- by-step.  Deep learning models can be supervised, semi-supervised, or unsupervised.  Advanced machine learning algorithms are used by tech giants, like Google, Microsoft, and Amazon to run entire systems and power things, like self driving cars and smart assistants. 30 15  Machine Learning vs. Deep Learning 31  In classical machine learning algorithms, feature extraction has to be done separately, while in deep learning models, it is done automatically within some layers in the model itself. 31 Feature Extraction 32  Feature extraction refers to the process of transforming raw data into numerical features that can be processed by the machine learning algorithms.  Feature extraction depends on the input and application.  Extracted features should be able to describe the samples in a way to help distinguishing the classes. 32 16  Intelligent Agents 33  An Intelligent Agent is an independent program or entity that interacts with its environment by perceiving its surroundings via sensors, then acting through actuators or effectors. An AI agent can have mental properties such as knowledge, belief, intention, etc.  Thus, an Agent runs in the cycle of perceiving, thinking, and acting. 33 Intelligent Agents 34 Important Definitions:  Sensors: devices that detect the change in the environment and send the information to other electronic devices. An agent observes its environment through sensors.  Actuators: component of machines that converts energy into motion. The actuators are only responsible for moving and controlling a system. An actuator can be an electric motor, gears, rails, etc.  Effectors: devices which affect the environment. Effectors can be legs, wheels, arms, fingers, wings, fins, and display screen. 34 17  Intelligent Agents 35  Examples of agents in general terms include:  Human agents: We are all intelligent agents! ◼ Have eyes, ears, and other organs that act as sensors. ◼ Have hands, legs, mouths, and other body parts act as actuators or effectors. ◼ Have ‘natural’ intelligence.  Robotic agents ◼ Have cameras and infrared range finders that act as sensors. ◼ Have various servos and motors performing as actuators. ◼ Have ‘artificial’ intelligence.  Software agents ◼ Have file contents, keystrokes as sensory input ◼ Acton those inputs displaying the output on a screen which acts as an effector ◼ Have ‘artificial’ intelligence, such as natural language processing (NLP) 35 Examples of Intelligent Agents in AI 36  Self-driving cars Environment Sensors Actuators and effectors Roads Camera Steering Other vehicles GPS Accelerator Road signs Speedometer, odometer, Brakes Pedestrians accelerometer, sonar. signal horn 36 18  Examples of Intelligent Agents in AI 37 Exercise: The following are examples of agents, find the environment, sensors, and actuators/effectors for each of them: 37 Examples of Intelligent Agents in AI 38  Cleaning Robot Environment Sensors Actuators and effectors Room Camera Wheels Table Dirt detection sensor Brushes Wood floor Cliff sensor Vacuum Extractor Carpet Bump Sensor Various obstacles Infrared Wall Sensor 38 19  Examples of Intelligent Agents in AI 39  Part-picking/ Bin-picking Robot: A technology in which a robot takes parts out of a container filled with unorganized, often irregularly shaped pieces and aligns them properly before sending them to the next station. Environment Sensors Actuators Conveyor belt Camera Jointed Arms Bins Joint angle sensors Hand/vacuum gripper 39 Learning Outcomes 40 Upon completion of the course, students will be able to: 1. Identify the importance of AI and Data Science for society 2. Perform data loading, preprocessing, summarization and visualization 3. Apply machine learning methods to solve basic regression and classification problems 4. Apply artificial neural networks to solve simple engineering problems 5. Implement basic data science and machine learning tasks using programming tools 40 20