Generative AI for Business with Microsoft Azure OpenAI

The outcome of this week is to understand foundational Machine Learning principles that enable Generative AI to perform tasks like creating new content, such as text and images, by learning from extensive datasets.

• Mathematical Foundations of Generative AI
• Understanding Machine Learning for Generative AI
• Connect NLP fundamentals with advanced Generative AI applications

The outcome of this week is to understand the Generative AI Landscape, fundamentals, and possibilities for businesses to solve problems and create products.

• Understanding Generative and Discriminative AI
• A brief timeline of Generative AI
• A peek into generative models
• Deconstructing the behavior of a large language models
• ML, DL, and GenAI applications in business
• Hands-on Demonstration of popular tools (ChatGPT & DALL-E)

• LLMs and the genesis of prompting
• A brief history of the GPT model series
• Accessing GPT through Azure
• Designing prompts for business use-cases using playground templates
• Prompting techniques (prompt templates, precise instructions and Chain-of-Thought Prompting)
• Ideating for prompts (prompt generation by induction and prompt paraphrasing)
• Understanding the concept of Prompt Engineering and its role in optimizing Azure OpenAI models' performance.
• Learning the capabilities of DALL-E in the Azure OpenAI service and use the DALL-E playground in Azure OpenAI Studio
• Introduction to Responsible AI

The outcome from this week is to get up to speed on the Python concepts that are needed to automate Prompt Engineering at scale, and understand the cost implications of using APIs.

• Speech recognition using Generative AI
• Whisper Architecture
• Audio transcription and translation
• Dual image-text understanding
• The CLIP model from OpenAI
• Image generation from text prompts
• Image generation architectures (GANs, Stable Diffusion)

The outcome from this week is to get proficient in using Prompt Flow to automate Prompt Engineering tasks, optimize workflows, and manage API integrations efficiently—all without coding, and understand the cost implications of using APIs

• Introduction to Prompt Flow
• Key features of Prompt Flow
• Setting up and integrating Prompt Flow with Azure AI Studio
• Navigating the Prompt Flow interface
• Exploring basic workflows in Prompt Flow
• Generating completions for prompts and manage model parameters 

This week focuses on applying Generative AI to multi-modal tasks, enhancing workflows that require a combination of text, audio, and visual data.

• Speech recognition using Generative AI
• Whisper Architecture
• Audio transcription and translation
• Dual image-text understanding
• The CLIP model from OpenAI
• Image generation from text prompts
• Image generation architectures (GANs, Stable Diffusion)

The outcome of this week is to learn how to use the Azure Open AI API key and the Python SDK to leverage Generative AI at scale for solving business problems

• Getting setup with your Azure Open AI key and Python SDK (for code track)
• Getting setup with your Azure Open AI key and Prompt Flow (for no-code)
• Completions and Chat API
• Kinds of APIs, Models, Token, Rate Limits and Pricing
• Evaluating Generative AI Outputs
• Generate completions to prompts and begin to manage model parameters
• Include clear instructions, request output composition, and use contextual content to improve the quality of the model's responses

The outcome of this week is to learn how to use Prompt Engineering to solve classification-type problems

• Text-to-Label Generation (Classification)
• Framing text classification tasks as Generative AI problem
• Sentiment classification
• Assigning themes to a body of text 
• Aspect-based sentiment analysis

• Content generation using Generative AI 
• Text-to-Text generation (Summarization)
• Abstractive summarization
• Evaluation metrics (GPT Similarity)
• Privacy protection in Generative AI
• Bias mitigation
• Managing Generative AI risks
• Security risks (Prompt Injection, Insecure Outputs, and Excessive Agency)

Problem Statement: The objective of this problem statement is to use aspect-based classification for sentiment analysis to identify the aspects of a product or service that customers are most satisfied with and those that need improvement. This will help businesses understand their customers better and make data-driven decisions to improve their products or services. By improving customer satisfaction and loyalty, businesses can increase customer retention rates, reduce churn rates, and ultimately increase revenue.

Identify characteristics of standard machine learning workloads, comprehend foundational principles of ML, and become acquainted with the prevalent machine learning methodologies.

• Identify regression, classification, and clustering machine learning scenarios
• Identify features and labels in a dataset for machine learning
• Describe the capabilities of automated machine learning
• Describe data and compute services for data science and machine learning
• Describe model management and deployment capabilities in Azure machine learning

Recognize various computer vision solution types and discover Azure tools for handling computer vision tasks.

• Identify common types of computer vision solution
• Identify features of optical character recognition solutions
• Capabilities of the Azure AI Vision service
• Capabilities of the Azure AI Face detection service

Identify features of typical NLP workload scenarios and explore Azure tools and services applicable to NLP workloads.

• Identify features and uses for key phrase extraction
• Identify features and uses for entity recognition
• Identify features and uses for language modeling
• Identify features of common NLP Workload Scenarios
• Identify Azure tools and services for NLP workloads

Focus on recognizing features of generative AI solutions and understanding the capabilities offered by the Azure OpenAI Service.

• Identify features of generative AI solutions
• Identify capabilities of Azure OpenAI Service