Fragments 2017

A conference on the mobile ecosystem in India

Mobile, AI and TensorFlow Lite

Submitted by Supriya Srivatsa (@sup95) on Jul 2, 2017

Section: Crisp Talk Technical level: Beginner Status: Awaiting details


Today we stand on the cusp of witnessing AI take over several aspects of our life. Google struck a chord when it announced at I/O that it was marking a shift from Mobile-first to AI-first. TensorFlow Lite allows for in device machine learning, without having to hit a remote server over the Internet each time. In this crisp talk, I will be speaking about the convergence of mobile and AI, and how TensorFlow Lite can be leveraged for the same.


Draft Outline:

  1. The convergence of Mobile and AI
    • Touch upon AI - past, present and future.
    • Why mix up Mobile and AI
  2. A little Background
    • A high-level overview of what neural networks are; what deep learning is.
    • How traditional machine learning with mobile works (Remote servers, hit an API)
  3. What is TensorFlow
    • What, why, how.
  4. What is TensorFlow Lite
    • How does it work behind the scenes.
    • Deep learning is compute intensive. How does Tensorflow Lite target mobile phones?
    • Packaging model within the app; the different ways of packaging app and model.
  5. Getting Started with TensorFlow Lite
    • Requirements
    • Configuration
  6. Applications and Scope of TensorFlow Lite
    • Look at some current applications
    • What Tensorflow Lite can help you achieve

Speaker bio

I am Supriya Srivatsa; I recently graduated from Amrita Vishwa Vidyapeetham and currently work at Xome as a software engineer - mobile. I love exploring and tinkering with new technology and have been deeply immersed in android app development lately. I have a keen interest in machine learning and also love participating in hackathons, breathing life into ideas to create a tangible product.


Preview video


{{ gettext('Login to leave a comment') }}

{{ gettext('You need to be a participant to comment.') }}

{{ formTitle }}
{{ gettext('Post a comment...') }}
{{ gettext('New comment') }}

{{ errorMsg }}