Session 1- Intuition, Learnings and Application.pptx

a intuitive presentation for ML beginners

1. GOOGLE DEVELOPER STUDENT CLUBS AI/ML: INTUITION, LEARNINGS and APPLICATION YADNYESH CHAKANE AI/ML Team – GDSC DIT

2. IMAGINE A NEURAL NETWORK...

3. Neural Networks are Function Approximators

4. First Principles Thinking

5. LINEAR REGRESSION – the blue line is the best fit line and the dots are data points

6. Gradient Descent

7. BUT WHAT WE ARE TRYING TO ACHIEVE HERE?

8. WE are trying to achieve profits by building AI models and products for the tech giants so that they can go up to the moon….. jk

9. Reinforcement Learning – Markov Decision Process

10. We are trying to solve complex problems using ML models like: - Protein Folding - Cancer and Tumor Detection using Image Segmentation - Creating Algorithms for Object Detection and 3D reconstruction in Self Driving Cars - Hyperspectral Image Analysis in Satellites - Drug Discovery using Large Language Models . . . . . . . . And many moreeeee ……..

11. Intuition The Cat:

12. semantic segmentation using U-Net Semantic segmentation of a satellite image

13. Tumor Analysis By Image Segmentation using U-Net Object Detection and Segmentation

14. Key Learnings : Intuition First Approach (First Principles Thinking) We need to identify problems first and build ML/DL Models around it AI/ML is full of Algorithms and Techniques to Improve them Dataset Preparation is the most important task in ML

15. THINK BIG, DO IT FAST - Sam Altman

16. How to optimize resources for better learning?

17. A Developer’s Approach

18. Some Good Practices

19. Learn in Public

20. RESOURCES Code written by other devs GitHub Repos Documentation Research Papers Medium Blogs Research Articles Research Paper Implementation through Code Twitter/X GitHub Pages of Researchers and Devs Exclusive Tutorials from Devs Create your own Resource Vault - links - notes - docs - papers - findings - observations

21. Arigatō