Plant AI: Project Showcase

Transcript

1. View Slide

2. View Slide

3. Motivation
   ● Read a disturbing headline
   

   View Slide

4. Motivation
   ● Read a disturbing headline
   Source: Times Of India
   

   View Slide

5. Motivation
   - Times Of India
   Most prominent Indian newspaper
   

   View Slide

6. Motivation
   ● Read a disturbing headline
   ● Researched a bit about the problem
   

   View Slide

7. Motivation
   ● Read a disturbing headline
   ● Researched a bit about the problem
   ● Crop losses
   Source: National Academy of Science, Singh et al.
   

   View Slide

8. Motivation
   ● Read a disturbing headline
   ● Researched a bit about the problem
   ● Crop losses
   ● We 🧡 tech and doing social good
   

   View Slide

9. The Problem
   ● Crop losses: diseases and deficiencies
   

   View Slide

10. View Slide

11. The Project
    ● Turns out almost all farmers have access to smartphones
    - Research Study by Hughes et al.
    

    View Slide

12. The Project
    ● Turns out almost all farmers have access to smartphones
    ● What does this do?
    ○ Uses AI to diagnose diseases from plant images early
    

    View Slide

13. The Project
    ● Turns out almost all farmers have access to smartphones
    ● What does this do?
    ○ Uses AI to diagnose diseases from plant images early
    ○ Provides actionable ways to solve diseases
    

    View Slide

14. The Project
    ● Turns out almost all farmers have access to smartphones
    ● What does this do?
    ○ Uses AI to diagnose diseases from plant images early
    ○ Provides actionable ways to solve diseases
    ○ Specific ways to solve
    

    View Slide

15. The Project
    ● Turns out almost all farmers have access to smartphones
    ● What does this do?
    ○ Uses AI to diagnose diseases from plant images early
    ○ Provides actionable ways to solve diseases
    ○ Specific and actionable ways to solve, AI also identifies species
    ○ Works Offline, PWA
    

    View Slide

16. The Project
    

    View Slide

17. The Project
    

    View Slide

18. The Project
    ● Turns out almost all farmers have access to smartphones
    ● What does this do?
    ○ Uses AI to diagnose diseases from plant images early
    ○ Provides actionable ways to solve diseases
    ○ Specific and actionable ways to solve, AI also identifies species
    ○ Works Offline
    ● Approximated decrease in plant losses: 67% -> 25%
    

    View Slide

19. Existing Solutions
    ● CropNet (Google), only Cassava plants
    

    View Slide

20. Existing Solutions
    ● CropNet (Google): only Cassava plants
    ● Plant Village (Hughess et al.): only a dataset
    

    View Slide

21. Existing Solutions
    ● CropNet (Google): only Cassava plants
    ● Plant Village (Hughess et al): only a dataset
    ● Plant Disease detector (Ramesh et al): unsatisfactory performance
    

    View Slide

22. Existing Solutions
    ● CropNet (Google): only Cassava plants
    ● Plant Village (Hughess et al): only a dataset
    ● Plant Disease detector (Ramesh et al): unsatisfactory performance
    ● Some other disease detectors: not optimized or usable on small
    devices
    

    View Slide

23. How we built this? - The Model
    ● Collected data
    

    View Slide

24. How we built this? - The Model
    ● Collected data
    ● Experimented with multiple architectures and hyperparameters
    

    View Slide

25. How we built this? - The Model
    ● Collected data
    ● Experimented with multiple architectures and hyperparameters
    

    View Slide

26. How we built this? - The Model
    ● Collected data
    ● Experimented with multiple architectures and hyperparameters
    ● Optimize the model to run on-device (in a PWA)
    

    View Slide

27. How we built this? - The Model
    ● Collected data
    ● Experimented with multiple architectures and hyperparameters
    ● Optimize the model to run on-device (in a PWA)
    ● Optimized the model to be just 12 MBs
    

    View Slide

28. How we built this? - The Model
    ● Collected data
    ● Experimented with multiple architectures and hyperparameters
    ● Optimize the model to run on-device (in a PWA)
    ● Optimized the model to be just 12 MBs
    

    View Slide

29. How we built this? - The Model
    ● Collected data
    ● Experimented with multiple architectures and hyperparameters
    ● Optimize the model to run on-device (in a PWA)
    ● Optimized the model to be just 12 MBs
    Wait, this is
    fast!💡
    

    View Slide

30. How we built this? - Web App
    ● Offline Support PWA
    

    View Slide

31. How we built this? - Web App
    ● Offline Support PWA
    ● Run the optimized model with TFJS
    

    View Slide

32. How we built this? - Web App
    ● Offline Support PWA
    ● Run the optimized model with TFJS
    ○ Individually fetch data flow graph and weights
    ○ Normalize images
    ○ Resize with nearest neighbour interpolation
    

    View Slide

33. How we built this? - Web App
    ● Offline Support PWA
    ● Run the optimized model with TFJS
    ○ Individually fetch data flow graph and weights
    ○ Normalize images
    ○ Resize with nearest neighbour interpolation
    ● Finally, deploy the web app to test out with real life scenarios🚀
    

    View Slide

34. Business Plan
    ● Farmers don't get charged
    

    View Slide

35. Business Plan
    ● Farmers don't get charged
    ● Potential plant product companies place ads
    

    View Slide

36. Business Plan
    ● Farmers don't get charged
    ● Potential plant product companies place ads
    ● After a quality check their ads appear under “How to Solve”
    

    View Slide

37. Business Plan
    ● Farmers don't get charged
    ● Potential plant product companies place ads
    ● After a quality check their ads appear under “How to Solve”
    ● Data, data and data
    

    View Slide

38. Business Plan
    ● Farmers don't get charged
    ● Potential plant product companies place ads
    ● After a quality check their ads appear under “How to Solve”
    ● Data, data and data
    ● Collect non sensitive but useful information
    

    View Slide

39. Business Plan
    ● Farmers don't get charged
    ● Potential plant product companies place ads
    ● After a quality check their ads appear under “How to Solve”
    ● Data, data and data
    ● Collect non sensitive but useful information
    ● Plant images are used anonymously to improve the ML
    

    View Slide

40. Business Plan
    ● Farmers don't get charged
    ● Potential plant product companies place ads
    ● After a quality check their ads appear under “How to Solve”
    ● Data, data and data
    ● Collect non sensitive but useful information
    ● Plant images are used anonymously to improve the ML
    ● Usage data is shared to plant product companies
    

    View Slide

41. Impact
    ● Can reduce crop losses by 67% -> 25%🚀
    ● Tested on real scenarios
    ● Featured on Microsoft Blog and YouTube🤗
    

    View Slide

42. Thank you!
    And we 🧡 open-source
    https://git.io/JujAg
    View on GitHub
    Research supported with Cloud TPUs from
    Google's TPU Research Cloud (TRC)
    

    View Slide