- The list of services offered by firebase - What is machine learning - Types of machine learning - How firebase helps to achieve ML with ML Kit - What is AutoML Vision Edge and how to use it to train model

1. Machine learning with firebase
ML Kit
Pankaj Rai
@raipankaj7

2. Agenda
Why firebase?
What is machine learning?
Machine learning with firebase

3. Why
Firebase? ● Available for Android, iOS and Web
● Reduce time for app development
● An all in one platform
● Helps to grow your business

6. Firebase
services
● Cloud Firestore - It’s a flexible, scalable
database for mobile, web, and server
development from Firebase and Google
Cloud Platform
● Firebase Realtime Database - It stores
JSON application data, like game state or
chat messages, and synchronizes
changes instantly across all connected
devices

7. Firebase
services
(cont)
● Firebase Remote Config - It stores
developer-specified key-value pairs to
change the behavior and appearance of
your app without requiring users to
download an update
● Firebase Hosting - It hosts the HTML,
CSS, and JavaScript for your website as
well as other developer-provided assets
like graphics, fonts, and icons

8. Machine Learning

9. Machine Learning
It is the learning in which
machine can learn by its own
without being explicitly
programmed. It is an
application of AI that provide
system the ability to
automatically learn and
improve from experience.
“A computer program is said to
learn from experience E with
respect to some task T and some
performance measure P , if its
performance on T , as measured
by P , improves with experience
E” - Tom Mitchell(1998)

10. Why machine
learning is
trending?
● Huge set of data availability
● Lots of computing power
● Effective ML algorithm

11. Machine
Learning is
for?
● Business leaders
● Software developer
● Data scientist

12. What is data
scientist?
● Understands statistics
● Knows ML softwares
● Having some problem domain

13. Types of
machine
learning
● Supervised Learning
● Unsupervised Learning
● Reinforcement Learning

14. Supervised
Learning

15. Unsupervised
Learning

16. Reinforcement
Learning

18. Capabilities Vision
Language
Custom

19. Vision ● Text recognition
● Face detection, Object detection and tracking
● Image labeling
● Barcode scanning
● Landmark recognition

20. Language
● Language ID
● On-device translation
● Smart reply

21. Custom
● AutoML Vision Edge
● TensorFlow Lite models

22. AutoML Vision Edge

23. ● Assemble training data
● Train model
● Evaluate model
● Publish or download model

24. Assemble
training data
● Upload labeled photos in zip folder
● No HD quality images required
● Supported formats: JPEG, PNG, GIF, BMP,
ICO
● Include multiple angles, resolutions, and
backgrounds for each label
● Get better result with least 10, and
preferably 100 or more photos

25. my_training_data.zip
|____accordion
| |____001.jpg
| |____002.jpg
| |____003.jpg
|____bass_guitar
| |____hofner.gif
| |____p-bass.png
|____clavier
|____well-tempered.jpg
|____well-tempered (1).jpg
|____well-tempered (2).jpg

26. Train model
● Create your first AutoML dataset or Add
another dataset
● Upload images or CSV files
● Verify training data and label the unlabeled
images
● Click on train model

28. Publish or
download
model
● By publishing update the model without
releasing a new app version
● Reduce the initial download size of your
app
● With publish and bundle it with your app,
the app will use the latest version available

29. Thank You
Pankaj Rai
https://twitter.com/raipankaj7