Self-Driving Apps - Using Machine Learning and AI to Improve App Performance

Explore machine learning and AI applications in app development through this 29-minute conference talk from Playtime San Francisco 2017. Discover real-world examples of developers successfully leveraging these technologies to enhance app and game performance. Dive deep into Instacart's case study, examining how they utilize machine learning for various aspects, including logistics and growth. Learn about the three-step workflow for training ML models, understand the value of machine learning in solving complex business problems, and gain insights into encoding techniques for product recommendations. Examine the challenges of the traveling salesman problem in the context of grocery picking, and get an introduction to TensorFlow architecture. Presented by Kevin Fives from Google Play and Montana Low from Instacart, this talk provides valuable insights for developers looking to integrate AI and machine learning into their applications.

Intro
Why Machine Learning? Can you write code that tells the difference between an apple and an orange?
You could write manual rules...
Creating manual rules requires lots of code
Machine Learning learns from examples
Workflow: 3 steps to train a ML model
ML can solve a variety of mission critical business problems
How to draw a Machine Learning model
Instacart value proposition
Four sided marketplace
Customer experience
Personal shopper experience
Encoding
New products
Competitive products
Recommended products
Unsupervised learning
Study the differences
Picking groceries
Let's get store data!
Traveling salesman problem
Results
Problem definition
Tensorflow architecture