Slide 1

Slide 1 text

Superpower your Android Apps with ML: Android 11 Rishit Dagli High School TEDx, TED-Ed Speaker rishit_dagli Rishit-dagli

Slide 2

Slide 2 text

● 11 Grade Student ● TEDx and Ted-Ed Speaker ● ♡ Hackathons and competitions ● ♡ Research ● My coordinates - www.rishit.tech $whoami rishit_dagli Rishit-dagli

Slide 3

Slide 3 text

● Sayak Paul (ML GDE) ● Khanh LeViet (Google) ● Hoi Lam (Google) Acknowledgements

Slide 4

Slide 4 text

● Mobile Devs looking for ways to build smarter apps ● Mobile Devs looking for ways to integrate ML in their existing apps easily Ideal Audience

Slide 5

Slide 5 text

Why care about ML in Android?

Slide 6

Slide 6 text

Why care about on-device ML in Android?

Slide 7

Slide 7 text

No content

Slide 8

Slide 8 text

No content

Slide 9

Slide 9 text

No content

Slide 10

Slide 10 text

No content

Slide 11

Slide 11 text

● Power Consumption Why should you care?

Slide 12

Slide 12 text

● Power Consumption ● Inference Time Why should you care?

Slide 13

Slide 13 text

● Power Consumption ● Inference Time ● Network availability Why should you care?

Slide 14

Slide 14 text

● Power Consumption ● Inference Time ● Network availability ● Privacy Why should you care?

Slide 15

Slide 15 text

ML Model Binding Plugin

Slide 16

Slide 16 text

Easier to use Enable Hardware acceleration Faster Development ML Model Binding Plugin What’s new for on-device ML in Android?

Slide 17

Slide 17 text

Importing a TF Lite Model

Slide 18

Slide 18 text

Importing a TF Lite Model

Slide 19

Slide 19 text

Importing a TF Lite Model

Slide 20

Slide 20 text

Using the TF Lite Model

Slide 21

Slide 21 text

Creating an Instance of model

Slide 22

Slide 22 text

Processing images

Slide 23

Slide 23 text

Passing in data

Slide 24

Slide 24 text

Passing in data

Slide 25

Slide 25 text

Passing in data

Slide 26

Slide 26 text

Passing in data

Slide 27

Slide 27 text

Adding labels

Slide 28

Slide 28 text

We are done

Slide 29

Slide 29 text

GPU Acceleration

Slide 30

Slide 30 text

A new ML Kit

Slide 31

Slide 31 text

Face detection Barcode scanning Image labeling Smart Reply Language Identification Vision Natural Language Object detection and tracking On-device Translation Text recognition Digital Ink Recognition Pose Detection N EW N EW

Slide 32

Slide 32 text

On-Device ML Better customizability Generic use cases A new ML Kit What does the latest ML Kit focus on?

Slide 33

Slide 33 text

Setup the model

Slide 34

Slide 34 text

Customize the model

Slide 35

Slide 35 text

No content

Slide 36

Slide 36 text

● High performance ● Super easy to use ● High customization too! TF Lite Model Maker

Slide 37

Slide 37 text

TF Hub tfhub.dev

Slide 38

Slide 38 text

TF Hub

Slide 39

Slide 39 text

bit.ly/a11-ml Demos!

Slide 40

Slide 40 text

Q & A rishit_dagli Rishit-dagli

Slide 41

Slide 41 text

Thank You rishit_dagli Rishit-dagli