What is AI?
AI is a broad field that encompasses the development of intelligent systems capable of performing tasks that typically require human intelligence, such as perception, reasoning, learning, problem-solving, and decision-making. AI serves as an umbrella term for various techniques and approaches, including machine learning, deep learning, and generative AI.
What is Machine Learning?
ML is subset of AI for understanding and building methods that make it possible for machines to learn. These methods use data to improve computer performance on a set of tasks.
What is Deep Learning?
Deep learning uses the concept of neurons and synapses similar to how our brain is wired.
What is Generative AI?
- Generative AI is a subset of deep learning because it can adapt models built using deep learning, but without retraining or fine tuning.
- Generative AI is capable of generating new data based on the patterns and structures learned from training data.
- It can create new content, including conversations, stories, images, videos, music, and code.
How can AI contribute to entertainment and media?
- Content generation: AI can create scripts, dialogues, or even complete stories for films, TV shows, and games.
- Virtual reality: AI can create immersive and interactive virtual environments for games or simulations.
- New generation: AI can generate articles or summaries based on raw data or events.
How can AI contribute to retail?
- Product review summaries: AI can generate review summaries for products so consumers can quickly find pertinent information.
- Pricing optimization: AI can model different pricing scenarios to determine optimal pricing strategies that maximize profits.
- Virtual try-ons: AI can generate virtual models of customers for virtual try-ons, which can improve the online shopping experience.
- Store layout optimization: AI can generate the most efficient store layouts to improve the customer shopping experience and boost sales.
How can AI contribute to healthcare?
- AWS HealthScribe: This AWS service empowers healthcare software vendors to build clinical applications that automatically generate clinical notes by analyzing patient-clinician conversations.
- Personalize medicine: By generating treatment plans based on a patient’s specific genetic makeup, lifestyle, and disease progression, AI can contribute to more effective, personalized care.
- Improve medical imaging: AI can enhance, reconstruct, or even generate medical images, like X-rays, MRIs, or CT scans, which can aid in better diagnosis.
How can AI contribute to life sciences?
- Drug discovery: AI can generate new potential molecular structures for drugs and accelerate the process of drug discovery and reducing costs.
- Protein folding prediction: AI can predict the 3D structures of proteins based on their amino acid sequence, which is crucial for understanding diseases and developing new therapies.
- Synthetic biology: AI can generate designs for synthetic biological systems, such as engineered organisms or biological circuits.
How can AI contribute to financial services?
- Fraud detection mechanisms: AI can help create synthetic datasets to improve AI and ML systems by simulating various money-laundering patterns.
- Portfolio management: AI can simulate various market scenarios and help in the creation and management of robust investment portfolios.
- Debt collection: AI can generate the most effective communication and negotiation strategies for debt collection to increase the rate of successful collections.
How can AI contribute to manufacturing?
- Predictive maintenance: By analyzing historical production data, AI can predict maintenance schedules that will provide the most efficient machine outputs and reduce downtimes.
- Process optimization: AI can generate the most efficient production processes by modeling different scenarios and optimizing for variables such as cost, time, resource usage, and so forth.
- Product design: AI can be used to create new product designs based on set parameters and constraints. It can generate multiple design options and optimize for factors like cost, materials, performance, and so forth.
- Material science: AI can help generate new material compositions with desired properties.
What is Computer Vision? Provide examples of AI applications that use CV and describe their business value.
Computer vision is a field of artificial intelligence that allows computers to interpret and understand digital images and videos. Deep learning has revolutionized computer vision by providing powerful techniques for tasks such as image classification, object detection, and image segmentation:
_Examples:_
- Auto manufactures can use computer vision technology to make self-driving cars safer and more reliable. Business value: Enhance customer experience
- Using computer vision in healthcare can improve the accuracy and speed of medical diagnoses, which leads to better treatment outcomes and increased life expectancy for patients. Business value: Improve business operations
- Computer vision image and facial recognition can swiftly identify unlawful entries or persons of interest, which fosters safer communities and works as a crime deterrent. Business value: Enhance customer experience
What is Natural Language Processing (NLP) Provide examples of AI applications that use CV and describe their business value.
NLP is a branch of artificial intelligence that deals with the interaction between computers and human languages. Deep learning has made significant strides in NLP. It can perform tasks such as text classification, sentiment analysis, machine translation, and language generation:
_Examples:_
- Insurance companies can use NLP to extract policy numbers, expiration dates, and other personal information. Business value: Sensitive data redaction
- Telecommunication companies use NLP to analyze customer text messages and suggest personalized recommendations. Business value: Customer engagement
- In the education industry, students use Q&A chatbots to address questions. Business value: Enhance student experience and engagement
What is Intelligent Document Processing (IDP) Provide examples of AI applications that use CV and describe their business value.
IDP is an application that extracts and classifies information from unstructured data, generates summaries, and provides actionable insights:
_Examples:_
- Financial services use IDP to extract important information from mortgage applications to accelerate customer response time. It also helps with the underwriting process by identifying incomplete loan packages, tax forms, pay stubs, and other missing data. Business value: Improve business operations, automation
- IDP, along with other applications such as optical character recognition (OCR) and NLP, helps eliminate the manual effort of processing documents such as contractual documents, agreements, court filings, and legal dockets. Business value: Improve business operations
- Using IDP in healthcare can help expedite business quickly and accurately by processing various document types, such as claims and doctor’s notes. Business value: Improve business operations
What is Fraud Detection? Provide examples of AI applications that use CV and describe their business value.
Fraud detection refers to the process of identifying and preventing fraudulent activities or unauthorized behavior with a system, process, or transaction:
_Examples:_
- Financial services use fraud detection for identity verification, payment fraud detection, transaction surveillance, and anti-money laundering (AML) sanctions. Business value: Improve business operations
- Fraud detection systems in the retail industry protect businesses from financial losses, safeguard customer accounts and data, and maintain trust and confidence in online transactions. Business value: Improve business operations
- The telecommunication industry uses fraud detection to identify any fraudulent activities in any of the following areas:
- Telecom
Roaming, premium rate service, and subscription fraud - Online
New account fraud, claims processing fraud, and promotion abuse - Retail
Credit card and online retail fraud
Business value: Improve business operations
When should ML and AI be used and for what kinds of problems?
- Coding the rules is challenging: Many human tasks cannot be solved properly using simple, rule-based solutions. Take spam filtering for instance. Determining whether an incoming email is legitimate or spam is a complex task that cannot always be effectively tackled through a set of predefined rules. There are many variables at play. When rules rely on too many factors, have overlaps, or need to be finely tuned, it becomes difficult for humans to code them accurately. ML can be used to effectively solve this kind of problem.
- Scale of the project is challenging: In the spam filtering example, a human might be able to look at a few hundred emails and decide if they are spam or not. However, scaling this task to scan through millions of emails would be tedious and inefficient. ML solutions are appropriate for large-scale problems like this.
What are the three ML types?
1 - Supervised Learning
2 - Unsupervised Learning
3 - Reinforcement Learning
What are some of the key factors to consider when selecting an appropriate generative AI model ?
- Model types
- Performance requirements
- Capabilities
- Constraints
- Compliance
List some models used in AWS and their functions and use cases:
A121 labs - Jurassic-2 Models
_Tasks_
Text generation
Summarization
Paraphrasing
Chat
Information extraction
_Use Cases_
Financial services – summarize lengthy documents
Retail – generate product descriptions
Amazon - Amazon Titan
_Tasks_
Text summarization
Classification
Open-ended Q&A
Information extraction
Embeddings
Search
_Use Cases_
Advertising – create studio quality images
Customer service – generate real-time abstract summaries
Anthropic - Claude
_Tasks_
Content generation
Text translation
Question answering
Text summarization
Code explanation and generation
_Use Cases_
Developer – code generation and debugging
Legal – parse legal documents and answer questions
Stability AI - Stable Diffusion
_Tasks_
Generate photo realistic images from text input
Improve quality of generated images
_Use Cases_
Gaming and metaverse – create characters, scenes, and worlds
Advertising and marketing – create ad campaigns and marketing assets
Cohere - Command
_Tasks_
Text generation
Information extraction
Question and answering
Summarization
_Use Cases_
Customer service – support chatbots
Retail – provide product descriptions
Healthcare – summarize key ideas from long text
Meta - Llama
_Tasks_
Question answering
Chat
Summarization
Paraphrasing
Sentiment analysis
Text generation
_Use Cases_
Customer service support – chatbots
What are some of the performance requirements to consider when considering a Gen AI model?
- Constraints
- Capabilities
- Compliance
- Cost
Name some business metrics for assessing the success of an AI application:
1 - User satisfaction
User satisfaction gathers user feedback to assess their satisfaction with the AI-generated content or recommendations.
Use case: Measuring and improving user satisfaction for an e-commerce website. An e-commerce company wants to monitor and enhance the overall user satisfaction with its website to increased customer loyalty, repeat purchases, and positive word-of-mouth.
2 - Average revenue per user
Average revenue per user (ARPU) calculates the average revenue generated per user or customer attributed to the generative AI application.
Use case: Analyzing and optimizing revenue generation per user. The marketing and product teams at an e-commerce company want to understand how effectively they are monetizing their user base and identify opportunities for improvement.
3 - Cross-domain performance
Cross-domain performance measures the generative AI model’s ability to perform effectively across different domains or industries.
Use case: Monitoring and optimizing a multidomain e-commerce platform. AnyCompany operates a large e-commerce platform with multiple domains catering to different product categories and geographic regions. They use the cross-domain performance metric to monitor the overall performance of their e-commerce platform across all domains.
4 - Conversion rate
Conversion rate monitors the conversion rate to generate content or recommend desired outcomes, such as purchases, sign-ups, or engagement metrics.
Use case: Optimizing an e-commerce website for higher conversion rates. A marketing manager for an online clothing store is responsible for analyzing and improving the website’s performance in terms of converting visitors into paying customers. To do this, they closely monitor the conversion rate metric, which measures the percentage of website visitors who complete a desired action, such as making a purchase.
5 - Efficiency
The efficiency metric evaluates the generative AI model’s efficiency in resource utilization, computation time, and scalability.
Use case: Improving production line efficiency. Example Corp Manufacturing Company operates a production line for assembling electronic devices. The company aims to optimize the efficiency of its production line to reduce costs and increase productivity.
Name the AWS AI/ML services stack:
- Amazon SageMaker: With SageMaker, you can build, train, and deploy ML models for any use case with fully managed infrastructure, tools, and workflows.
- Amazon Comprehend: Amazon Comprehend uses ML and natural language processing (NLP) to help you uncover the insights and relationships in your unstructured data.
- Amazon Translate: Amazon Translate is a neural machine translation service that delivers fast, high-quality, and affordable language translation.
- Amazon Textract: Amazon Textract is a service that automatically extracts text and data from scanned documents.
- Amazon Lex: Amazon Lex is a fully managed AI service to design, build, test, and deploy conversational interfaces into any application using voice and text.
- Amazon Polly: Amazon Polly is a service that turns text into lifelike speech.
- Amazon Transcribe: Amazon Transcribe is an automatic speech recognition (ASR) service for automatically converting speech to text.
- Amazon Rekognition: Amazon Rekognition facilitates adding image and video analysis to your applications. It uses proven, highly scalable, deep learning technology that requires no ML expertise to use.
- Amazon Kendra: Amazon Kendra is an intelligent search service powered by ML. Amazon Kendra reimagines enterprise search for your websites and applications.
- Amazon Personalize: Amazon Personalize is an ML service that developers can use to create individualized recommendations for customers who use their applications.
- AWS DeepRacer: AWS DeepRacer is a 1/18th scale race car that gives you an interesting and fun way to get started with reinforcement learning (RL).
