Skip to main content

OpenJS World 2021 Keynote Recap: Building Great Web Experiences with AMP and TensorFlow

By October 5, 2021Blog, OpenJS World

During the OpenJS World Keynote Panel, Jeffrey Jose and Sandeep Gupta spoke on the best practices to follow to create great web experiences and the importance of machine learning to create an interactive and communicative web app.

As user expectations have evolved, web developers have a greater responsibility of evolving with them to create great unique web experiences. 

Jeffrey Jose, Product Manager at Google working on AMP, spoke with Sandeep Gupta, Product Manager on TensorFlow at Google. This keynote session is divided into sections. The first part takes a look at how page experience and AMP work together. Then Gupta explains the use of Tensorflow.js to harness the power of machine learning to build novel experiences for the web.

Jose goes on to explain that a useful way of understanding User Experience is by using the four UX Pillars: Loading, User Annoyance, Security & Privacy, and Accessibility. Additionally, the core of vitals are not just a set of metrics but also a set of threshold guidance that map to user expectations. To further illustrate this, Jose gives the example of how the Chrome team has done a lot of research to come up with guidance to create a metric of performance.

Gupta emphasizes that Machine Learning touches our lives daily as it is spread across multiple fields like healthcare and education. It gives people new ways of interacting. An example of how Machine Learning is improving web experience and communication is how L’oreal uses it for a virtual make-up try on experience.

In their concluding thoughts, they encourage other users to continue following best practices for creating a better web experience. Machine Learning is an important component of this and gives your web application “superpowers.” 

Full video here

Broken down by section:

  • Speaker introduction 0:02
  • The Web 0:28
  • UX Pillars 1:16
  • Page Experience 2:13
  • PX Signals 2:26
  • Thresholds 4:58
  • ML and Web Experience 9:26
  • Does this mean you must learn Python? 12:16
  • An example of text search using the Q & A model 14:46
  • Object Recognition 16:58
  • Train your own custom models 18:03