1 AI Fundamentals from Slides

• Artificial intelligence capability of computer systems to imitate intelligent human behavior
• AI is a branch of computer science dealing with the simulation of intelligent behavior in computers
• The term "artificial intelligence" was first coined in 1956 by computer scientist John McCarthy at a Dartmouth Conference
• AI can mimic elements of human behavior to help make better decisions
• AI is not meant to replace humans, but augment their abilities
• AI can enhance decision-making, customer service, content generation, healthcare, and education

Historical Perspective

• AI dates back to the 1950s with many branches and approaches having developed since
• ChatGPT and other solutions is built on previous decades of work

Computer Vision

• AI technology that mimics human vision by analyzing images and videos
• This technology automates visual tasks, image-based decision-making, and problem-solving
• It assists in healthcare by analyzing medical images like MRIs
• It helps autonomous vehicles detect objects, pedestrians, and traffic signals
• It ensures quality control in manufacturing by detecting defects

Expert Systems

• AI applications designed to emulate human expert decision-making through specialized knowledge and reasoning
• Customer service expert systems help schedule orders, route requests, and solve problems
• Telecommunications expert systems make decisions about network technologies and maintenance
• Mechanical engineering expert systems troubleshoot complex electromechanical machinery designs
• Expert systems require extensive manual coding

Machine Learning

• This allows computer systems to enhance their performance by learning from data
• ML uses featurized data instead of explicit programming for tasks
• ML developed from expert systems
• Machine learning can be used in healthcare with personalized medicine, retail with product recommendations, and recruiting with automated job boards

Neural Networks

• A type of machine learning algorithm inspired by the human brain with interconnected nodes that learn to recognize patterns by training
• Used to overcome the limitations of traditional machine learning by automatically learning complex patterns from raw, high-dimensional, and unstructured data
• Healthcare utilizes neural networks to enhance medical image analysis, diagnosis, drug discovery, and personalized treatments
• Financial services utilizes neural networks to detect fraud, assess risk, assist in trading, and predict customer behavior
• Robotics utilizes neural networks to enable advanced perception, motion planning, and adaptive control for intelligent autonomous systems

Large Language Models and Generative AI

• LLMs and generative Al are creative AI branches making new things such as: text, images, music
• Made up of deep learning, neural networks and NLP

Large Language Models

• Al systems designed to understand, generate, and assess human language using large text datasets
• LLMs are used in copywriting to improve original style and voice by systems like Al21 Wordspice
• LLMs can produce code such as Amazon CodeWhisperer, including SQL queries
• LLMs can generate text for documentation or write a short story like Alexa Create
• LLMs play a role in creating natural language products and services focused on text generation and analysis

Generative AI

• Involves deep-learning models capable of producing high-quality text, images, and other content
• Generative AI fosters innovation by automating creativity, customizing user experiences, and creating new content
• In conversational Al, programs like ChatGPT generates answers, assists with tasks, and creates text content
• In data generation, programs such as Mostly Al creates fully anonymous synthetic data from original data
• In media creation, programs such as SORA produces audio, video, and multimedia content

Fine-Tuning

• Customizes an Al model with specific data to improve domain-specific tasks
• Fine-tuning is used in media and marketing to generate specialized, creative content
• Fine-tuning is used to tailor customer support responses

Retrieval Augmented Generation

• Extends an Al model by retrieving real-time external data without modifying it
• Useful in legal and medical fields to ensure accuracy and provide factual information
• Useful in quickly changing environments by accessing the latest data

Agentic AI

• Agentic Al is a sophisticated form of artificial intelligence to autonomously perform tasks, make decisions, and interact with users using its acquired knowledge
• Agentic Al combines LLMs, NLP, and Rules-Based Systems and acts as an independent decision-making "agent"
• Agentic Al solves problems by making decisions in healthcare, finance, and autonomous vehicles
• Agentic AI adapts dynamically operating within accountability guardrails

"Disruptive Innovation"

• An innovation that upends markets by offering simpler, cheaper, or more accessible solutions, shaking up the status quo
• When newcomers outsmart giants with fresh thinking that provides a unique value, reshaping thinking, and how progress is defined

Timeline creating value

• 1990 - The Web enabled brands to interact with customers through HTML and e-commerce
• Mobile technology made the internet everywhere and provided intimate experiences
• Algorithms and personalization techniques were used to target advertising and influence customers
• Blockchain technology gave rise to Bitcoin, cryptocurrency, and NFTs for security and authentication
• LLMs. Deep Learning, NLP.This is why Generative AL exists today

Examples of AI

• Al reduces credit card fraud by efficiently examining billions of transactions per day with Bank of America
• Duolingo uses AI to speed, provide answer explanations, and have conversations
• Apple fitness used AI to access health coaches
• Walmart uses AI to fulfill consumer needs