Building AI Applications with Microsoft Semantic Kernel

Building AI Applications with Microsoft Semantic Kernel
Contributors
About the author
About the reviewer
Preface
   Who this book is for
   What this book covers
   To get the most out of this book
   Download the example code files
   Conventions used
   Get in touch
   Share Your Thoughts
   Download a free PDF copy of this book
Part 1:Introduction to Generative AI and Microsoft Semantic Kernel
1
Introducing Microsoft Semantic Kernel
   Technical requirements
      Obtaining an OpenAI API key
      Obtaining an Azure OpenAI API key
   Generative AI and how to use it
      Text generation models
      Understanding the difference between applications and models
      Generating text using consumer applications
      Generating images
   Microsoft Semantic Kernel
      Installing the Microsoft Semantic Kernel package
   Using Semantic Kernel to connect to AI services
      Connecting to OpenAI Services using Python
      Connecting to OpenAI services using C#
   Running a simple prompt
      Running a simple prompt in Python
      Running a simple prompt in C#
   Using generative AI to solve simple problems
      Creating semantic functions
      Creating native functions
   Plugins
      The config.json file for the knock-knock joke function
      The skprompt.txt file for the knock-knock joke function
      The config.json file for the semantic function that explains jokes
      The skprompt.txt file for the explain joke function
      Loading the plugin from a directory into the kernel
   Using a planner to run a multistep task
      Calling the Function Calling Stepwise planner with Python
   Summary
   References
2
Creating Better Prompts
   Technical requirements
   A simple plugin template
      The skprompt.txt file
      The config.json file
      Calling the plugin from Python
      Calling the plugin from C#
      Results
   Improving the prompt to get better results
      Revising the skprompt.txt file
      The result
   Prompts with multiple variables
      Requesting a complex itinerary with Python
      Requesting a complex itinerary with C#
      The result of the complex itinerary
   Issues when answering math problems
   Multistage prompts
      CoT – “Let’s think step by step”
      Implementing CoT with Python
      Implementing CoT with C#
      Results for CoT
      An ensemble of answers
   Summary
   References
Part 2: Creating AI Applications with Semantic Kernel
3
Extending Semantic Kernel
   Technical requirements
   Getting to know the core plugins
      An example – Using the TimePlugin
   Introducing the application – Validating grants
      Directory structure of our application
   Developing native plugins
      The directory structure of our plugins
      Checking the structure of our Excel spreadsheet
      Additional checks
      Evaluating the Word document
   Developing semantic plugins
      Evaluating the grant proposal with a semantic plugin
   Summary
4
Performing Complex Actions by Chaining Functions
   Technical requirements
   Creating a native plugin that generates images
      Writing a DALL-E 3 wrapper in Python
      Writing a DALL-E 3 wrapper in C#
   Using multiple steps to solve a problem
      Generating an image from a clue
      Chaining semantic and native functions with C#
      Chaining semantic and native functions with Python
   Dealing with larger, more complex chains
      Preparing our directory structure
      Understanding the flow of our process
      Creating the native function to process a folder
      Modifying the Excel native plugin
      Modifying the Word native plugin
      Modifying the semantic functions
   Creating and calling the pipeline
   Summary
   References
5
Programming with Planners
   Technical requirements
   What is a planner?
   When to use a planner
   Instantiating a planner
   Creating and running a plan
      An example of how a planner can help
   How do planners work?
   Controlling home automation with the planner
      Creating the native functions
      Adding a semantic function to suggest movies
      Invoking the planner
   Summary
6
Adding Memories to Your AI Application
   Technical requirements
   Defining memory and embeddings
      How does semantic memory work?
      Embeddings in action
   Using memory within chats and LLMs
      Using memory with Microsoft Semantic Kernel
      Using memory in chats
      Reducing history size with summarization
   Summary
Part 3: Real-World Use Cases
7
Real-World Use Case – Retrieval-Augmented Generation
   Technical requirements
   Why would you need to customize GPT models?
   Retrieval-augmented generation
      Creating an index
      Uploading documents to the index
      Using the index to find academic articles
      Using RAG to create a summary of several articles on a topic
   Summary
   References
8
Real-World Use Case – Making Your Application Available on ChatGPT
   Technical requirements
   Custom GPT agents
      Creating a custom GPT
      Creating a custom GPT that supports actions
      Creating a web API wrapper for the native function
      Deploying your application to an Azure Web App
      Connecting the custom GPT with your custom GPT action
   Summary
Index
   Why subscribe?
Other Books You May Enjoy
   Packt is searching for authors like you
   Share Your Thoughts
   Download a free PDF copy of this book