Generative AI as a thought partner

An overview of GenAI, Prompt Engineering and GitHub Copilot.

Also includes some demos.

1. Generative AI as a thought partner Nimesh Desai 9 August 2024

2. Understanding Generative AI

3. Source:  PWC

4. Source:  SMB Group

5. Applications and Use Cases Image and Video Generation Generative AI can be used to create realistic images and videos with applications in gaming, film, and virtual reality. Text Generation Generative AI can produce natural language text, including news articles, product descriptions, and chatbot conversations with potential applications in content creation, customer service, and marketing. Medical Applications Generative AI can assist medical professionals with image analysis, diagnostic support, and drug discovery, enabling better patient outcomes and more effective treatments. Software Development Generative AI can help generate code, algorithms, documentation, comments, test cases and automation. It can also help in bug identification and fixes. AI can help generate design prototypes, review code and setup CI/CD pipeline.

6. Exploring Prompt Engineering

7. The Role of Prompts in AI Models What are Prompts? Prompts are inputs that guide AI models to generate specific outputs. Importance of Prompts in AI Models Prompts are important in AI models because they help to improve the quality and accuracy of generated outputs by directing the model's attention towards specific tasks. Art of communicating eloquently with AI Basic prompt –Getting the whole library Advanced prompt – Getting a specific book

8. Techniques for Effective Prompt Engineering Clear and Specific Prompts Effective prompts are clear, specific, and relevant to the desired output. They should provide enough information to guide the user towards the desired action. Prompt Tuning Prompt tuning involves adjusting prompts to achieve the desired output. This technique involves testing different prompts and adjusting them based on the results. Prompt Programming Prompt programming involves designing prompts using programming languages. This technique allows for more flexibility and customization in prompt design.

9. Advanced Prompt Engineering Advanced prompt engineering involves creating more complex prompts to guide AI models to generate specific outputs. 0-shot prompting – no examples. Few-shots prompting – with specific examples. Chain-of-thought prompting - complex reasoning capabilities through intermediate reasoning steps. Tree-of-thought prompting - more complex version of chain-of-thought prompting that involves creating a branching tree of prompts to guide the AI model (Panel Discussion) Chain-of-density prompting - technique that involves creating prompts of varying densities to guide the AI model towards more complex and nuanced outputs. Prompts demo - https://github.com/LinkedInLearning/advanced-prompt-engineering-techniques-3817061

10. Introduction to GitHub Copilot

11. Features and Capabilities Autocompletion GitHub Copilot has an advanced autocompletion feature that suggests code snippets to complete your code as you type. It uses machine learning to predict the code you are trying to write, which can improve coding efficiency. Debugging GitHub Copilot has a built-in debugging feature that helps identify and fix errors in your code. It provides real-time feedback on code execution, which can speed up the debugging process. Code Generation GitHub Copilot can generate code snippets based on natural language descriptions of what you want to achieve. This feature can save time and reduce the need for manual coding.

12. Real-World Applications and Examples Automated Code Suggestions GitHub Copilot provides automated code suggestions to developers, allowing them to write code more efficiently and accurately, which ultimately improves the coding process and boosts productivity. Code Completion GitHub Copilot can provide code completion suggestions based on the context of the code being written, helping developers save time and write code more accurately. Code Optimization GitHub Copilot helps developers optimize their code by suggesting changes that can improve performance, reduce complexity, and make the codebase more maintainable.

13. Different AI Tools

14. Demo

15. Generative AI as a thought partner Initial Data: Stories about Indian heroes in English language Project: Create a simple online quiz game to create awareness about these heroes. Create an online blogsite depicting the stories

16. Heroes Quiz Question (random sequence) Multiple Choices (random order) Hints (random order) Scoring Data handling Desktop/Mobile/Tablet compatible Deployment Keep data in online DB Hints Options Front Answer + Description Score Back

17. Quiz Development Code generation using ChatGPT Options/Hints/Description Generation Technologies – JavaScript, CSS, HTML, MongoDB Questions - JSON Hosting – GitHub/Netlify Database (to store users) – MongoDB ( https://cloud.mongodb.com/ ) Link - https://indianheroesquiz.netlify.app/

18. Indian Icons Static Blog Site Show stories of Indian icons Categories Menu Readable Layout Story of the day References Desktop/Mobile/Tablet compatible Home Story Heading Intro Story of the Day Contact Name Story References

19. Blog Development Technologies – Jekyll, Ruby, Minima Hosting – GitHub/Netlify Link - https://indianicons.in/

20. Project Ideas

21. Project Ideas Micro-biography Text-based/Graphical summary of one’s personal/professional journey Community Library Share/sell/buy used books, book-based events (reading sessions, activities, …) Sentiment Analysis Sentiment analysis of text data and generate sentiment scores Company’s internal employee feedback Perception about a particular product/movie etc